ScreenDataSharingLevelUpdater
- change names
- fixup logic
- implement EmailService

Create a command line tool and a service to expire screens where the expirationdate > dateArg.  (set ScreenDataSharingLevel to MUTUAL_SCREENS)

Create a command line tool and a service to expire screens where the expirationdate > dateArg.  (set ScreenDataSharingLevel to MUTUAL_SCREENS)

code cleaning

[#2119] Optimize ScreenResultParser for scalability with large import sets
- simplify tests!

Add update of the Screen.dataPrivacyExpirationDate on parse and load

Fixed model unit tests that test adding/removing elements from a "collection of elements"

[#2119] Optimize ScreenResultParser for scalability with large import sets
- finish tests
- fix tx annotation on ScreenResultLoader methods
- cleanup & testing

UserViewer and ScreenViewer ignore redundant set<Entity>() calls, eliminating duplicate warning messages, and putatively improving UI performance

clean up logic around append/delete/throw error on existing ScreenResult

[#2119] Optimize ScreenResultParser for scalability with large import sets
- add explicit AssayWell delete on the deletion of ScreenResult: memory heap issues were encountered when deleting the SR without doing this.

fix (temporary?) work-around for ignore-duplicates: now the first duplicate found will be the one recorded, second is ignored if flag is set

error if SR exists

[#2119] Optimize ScreenResultParser for scalability with large import sets
- remove DataModelViolation on Library WellType != Assay WellType

clean up logic around append/delete/throw error on existing ScreenResult

accidental check-in, removing

[#2040] auto-add screener facility usage roles when user associated with a screen

Enhancements and fixes to UserAgreementUpdater, including:
- Add screensaverUser role on update
- Allow update to proceed even if user already has a DSL specified (allows it be upgraded/downgraded)
- Handle misconfigured database (missing checklist item, attached file type) more gracefully
- Record update activities independently, and log old values
- Show current DSL for user and lab head, and whether user has screensaverUser role
- Fix issuance of warnings for User and Screen mismatched DSLs, which did not trigger when scrolling through data table in entity viewing mode

[#2119] Optimize ScreenResultParser for scalability with large import sets
- refactor ScreenResultImporters to use ScreenResultLoader
- TODO: is there a memory delete problem with AssayWells?, test UI, ignoreWellType issue

commented out code in ScreenResultParser that was effectively a no-op

added TODO comments

- Issue warning to admins when viewing user that have a small molecule data sharing level that is different than their lab head's data sharing level.
- Display data sharing levels in warning message about mismatched screen DSLs, and only issue if screen has a lab head

corrected user role value

Issue warning to admins when viewing small molecule screens that have a data sharing level that is too restrictive w.r.t. to lab head's data sharing level

reinstated WellSearchResults "assay well type" column (which was taken out with the intention of using AssayWay.assayWellType, but will have to wait until the ResultValue<->AssayWell reln is implemented)

implement persistence test for SRP, refactor ScreenResultImporter for testing

[#2119] Optimize ScreenResultParser for scalability with large import sets
- changed to better expose assay well creation problems
- use incrementalFlush flag to determine to clear SR wells, assaywells

[#2119] Optimize ScreenResultParser for scalability with large import sets
- simplify query, let postgresql optimize.  This version is actually faster.

[#2119] Optimize ScreenResultParser for scalability with large import sets
- clean up debug code

- WellSearchResults now disables sorting and searching on data columns that
only used to share "mutual positives".  These columns must not be
sortable or searchable, to avoid a prying user from inferring hidden values 
in these data columns (for performance reasons, the hidden data cannot be
filtered at the database level, and thus exists client-side, and is only
hidden at the highest UI level).

- Added TableColumn.isSortable()

- Added TableColumnManager.getSortableSearchableColumns(), which only contains
the subset of columns where TableColumn.isSortable() returns true

- Renamed HasFetchedPaths.isSingleProperty, to isFetchableProperty, which
better describes the intent

- Updated WellSearchResultsTest to run all tests with both InMemory and
VirtualPagingDataModels, now that WSR can be configured either way, regardless
of dataset size.

[#2119] Optimize ScreenResultParser for scalability with large import sets
- update the populate insert

[#2119] Optimize ScreenResultParser for scalability
- clean up and remove debug

[#2119] Optimize ScreenResultParser for scalability with large import sets

- workaround code to populate screen_result_well_link using SQL rather than HQL.

Added FetchedPaths.isSingleProperty() operation.  This will allow data tables to support unsortable and unsearchable columns, which is useful when using VirtualPagingDataModel that has columns derived from multiple entity properties.

EntitySearchResults no longer attempts to dynamically determine whether to use InMemoryDataModel or VirtualPagingDataModel.  Instead, defaults to InMemoryDataModel, and allows concrete subclasses override the behavior, as desired, in buildDataTableModel().

fixed NPE in LibraryViewer

fixed spring config file for SRP cmd-line app

Users can view Lab Activities, CPRs, and Status Items of their own screens, but not those of others' screens

Allow users to see their own CPRs via "Browse CPRs" command

Renamed data access roles

Users can now see their own data access roles and facility usage roles

- Screen data sharing level property is now visible to all users that can view the screen
- Screen data privacy expiration date property is now admin-only, and added as a new column to Screen Search Results

UI label rename

Renamed "published" Screen data sharing level to "public"

renamed user agreement checklist items and attached file type

no longer updating screen.data_privacy_expiration_date in 3614.sql (need to manually determine this value for many screens)

updated 3614.sql to default screens to DSL 3

fixes, updates to migration scripts

added migration script for DSL-related upates for ICCB-L production database

ChecklistItemEventComparator guarantees non-equality for different ChecklistItemEvent entities, avoiding clashes in ordered sets.  Fixes bug whereby migrated CIEs that ended up with same creation date caused "lost" CIEs in UI (and data loss on save).

renamed private data member

fixed migration script to set checklist_item_event performed_by and created_by fields

WDAP now calculates restrictions on reagents without having to navigate to the associated library.  This fixes a LazyInitEx when selecting reagent columns in WellSearchResults, without also selecting library columns.

changed display name for screensaverUserRole to "Screensaver User Login"

fixed migration script to preserve existing rnaiScreener data access roles

simplified schema migration script that updated screen data sharing levels, such that it should work for 3rd parties; ICCB-L will need a much different, custom migration script

fixed typo

fixed NPE is WellSearchResults

fixed NPE in debug output code in WDAP

WellSearchResults now respects DataAccessPolicy for ResultValuess and ResultValueTypes

minor updates

Reverted some of the recent cascade changes that were made to prevent AssayWells and ResultValues from being updated when a Screen is reattached to a Hibernate session (in particular, via a Hibernate merge operation).  Merge cascade remains elided, in order to address performance issues when users edit/save screens.  Reinstated save/update/persist, which broke many unit tests.

added debug output to WebDataAcessPolicy, to help verify proper granting of screen access permissions

- Fixed problem in how new roles were added to a user in UserAgreementUpdater (implied roles were not being added)
- Also, changed checklist item name to match value in ICCB-L database

fixed web.xml security constraints, now that generic 'screener' role no longer exists; now allowing screens/*, screens/screenresults/* and /cherrypickrequests to be accessible to users with login privileges (the DataAccessPolicy provides the necessary protections)

cascade changes to prevent AssayWells and ResultValues from being updated when a Screen is reattached to a Hibernate session (in particular, via a Hibernate merge operation)

fixed problem with saving user after editing checklist items; required reloading of admin user in AbstractBackingBean.recordUpdateActivity()

fixed Hibernate mapping for ChecklistItemEvent.updateActivities

Added UserAgreementUpdater service and UI page, accessible from UserViewer.

fixed typo in migration script

Rename ScreeningRoomUserFacilityUsage to FacilityUsageRole, with similar renaming throughout UI.

Revert removal of ScreenResult.shareable property, which we may still need to use for screens that are published.

Revert removal of ScreenResult.shareable property, which we may still need to use for screens that are published.

Added confirmation prompt on deleting a user role.

removed commented-code for "Browse Screeners" menu command (which no longer exists)

- For users with no roles, show "<not yet specified>" in UserSearchResult and UserViewer

- Added 'userRolesAdmin' role
- Only 'userRolesAdmin' admins can edit user roles in UserViewer

- Extracted appropriate roles from ScreensaverUserRole into ScreeningRoomUserFacilityUsage
- Added ScreeningRoomUser.facilityUsages collection
- Update UserViewer and UserSearchResults to show new facilityUsages data
- Removed the implied "screener" role from ScreensaverUserRole
- Removed 'Browse Screeners' command from main menu

In EntityUpdateSearchResults, "date created" column is now a DateTime column, allowing sorting to respect time-of-day.

added DateTime column

hide "Update History" panel when in edit mode

- Made "entity update history" UI panel available in all entity viewers (provided they have a non-null EntityUpdateSearchResult).  This now includes UserViewer and ScreenDetailViewer.
- AbstractEditableBackingBean.recordUpdate() now outputs "save" prefix
- Added a DateTimeConverter 

- Added "Update History" panel and data table to User Viewer
- Refactored "isPanelCollapsed" properties of backing beans, by pulling getIsPanelCollapsed() method up into AbstractBackingBean, and also updated older property-per-panel design in some beans.

- Removed (recently added) "Data Sharing Level" UI section from User Viewer, after realizing that we could accommodate our UI requirements by managing the data sharing level roles within the existing "Role" UI section.  This is made possible by treating user top-level, extant roles as "primary" roles and only showing the primary roles in the user's roles table.  Furthermore, we do not provide the roles that imply a user's extant primary roles.  In this way, primary roles appear as mutually exclusive choices, as desired.
- Only changes to primary roles are recorded in the audit log for ScreensaverUser.
- Added get{Primary,Implied}ScreensaverUserRoles() methods to ScreensaverUser.
- User Search Results now only shows primary roles in the roles column.

added "Data Sharing Level" column to ScreenSearchResults

AuditedAbstractEntity.dateCreated property changed to "not null"

fixed bug where clicking on name in Users Browser was broken (no action occurred)

fix for User Viewer adding of audit log entries on user role changes

- Update constructors in ScreensaverUser and ScreeningRoomUser to allow new, uninitialized entity (for UI-based creation) to set created by and create date properties
- Fixed UI problems on saving new user and screen entities.
- User Viewer adds audit log entries on user role changes.

fixed bug where screen data sharing level prop not being persisted after edited from the UI

fixed bug in model testing framework where getter/setter persistence test was not actually testing the persisted value, but rather only the in-memory value

fixed reset problem in ScreenDetailViewer

made dataAccessPolicy spring bean to be of scope "session"

renamed *_update_activities schema tables to *_update_activity, for consistency with schema naming conventions

- Fixed recently broken UI for checklist items
- ChecklistItemEvent is now an immutable entity, as it should have been originally

Screen.dataSharingLevel mapped to schema as int instead of text, since text value is really for UI display, and too verbose for db storage

- Added data sharing levels and data privacy expiration date properties to Screen Viewer UI
- Updated text values of ScreenDataSharingLevel enum
- Added NullSafeUtils.toString() methods

fixed NPE on update of data sharing level in UserViewer

use of ss:property adminOnly attribute userViewerBody.xhtml

- Added "Small Molecule Data Sharing Level" property to UserViewer.  The
various roles for small molecule data sharing levels are no longer visible in
the generic "Roles" property UI, and are now only configured via this new data
sharing level property UI.  This separation was necessary since they are
mutually exclusive on selection, and also have implied roles, which would have
made for a confusing UI if they stayed in the "Roles" UI.

- AbstractEditableBackingBean now manages update comments and provides
recordUpdateActivity() methods to update the audit log of any
AuditableAbstractEntity.

- Added optional "update comments" filed to editSaveCancel.xhtml, to support
updates to AuditableAbstractEntity audit logs (experimental).

- Removed in-memory updates of ScreensaverUser.activitiesPerformed() due to
Hibernate cascade/merge problems when adding update admin activities to user
(we never rely upon this in-memory relationship direction anyway)

fixed eager fetching of Screen.pinTransferApprovalActivity in ScreenViewer

- added missing Hibernate cascade annotations to getUpdateActivities() methods

- added missing Hibernate cascade annotations to getUpdateActivities() methods

removed final remnants of ChecklistItemEvent.entryActivity property

fixes for web app initialization

added SpringWebContextTest, to identify problems in Spring web context config files.

- Merged in 1.8.1-dev branch:
  merge -r3500:HEAD http://forge.abcd.harvard.edu/svn/screensaver/branches/iccbl/1.8.1-dev
- Note: merge of ScreenResultParser was non-trivial and is still untested, and memory optimizations may have regressed, due to introduction of AssayWell

Added Screen.dataPrivacyExpirationDate property

fixed imports

fixed schema migration script

[#2104] fix loading failures for screen results (iccbl_support)
addressing
[#2119] Optimize ScreenResultParser for scalability with large import sets
- also clear out the wells, gives another ~100MB of free memory
- Do not cache the ResultValues on the ResultValueType objects during parsing

revert the validateAssayWellType log message to warn level

[#2104] fix loading failures for screen results (iccbl_support)
addressing
[#2119] Optimize ScreenResultParser for scalability with large import sets
- also clear out the wells, gives another ~100MB of free memory

schema update script for administrative_activity table changes and new <entity>_update_activity tables (audit logs)

- LibraryContentsVersions is no longer an AuditedAbstractEntity, since the creation properties were redundant with the loading activity.
- Corrected unit test class for LibraryContentsVersion, which was ineffectual
- Corrected LCV entity ID setter method visibility
- Corrected model unit test output

Replaced ChecklistItemEvent.entryActivity with similar properties that are now automatically provided by the new AuditedAbstractEntity superclass.

updated javadocs

- Model enhancements to more easily accommodate auditing information for
field-level data changes, which will be need be used in a couple ways for the
Data Sharing Level release.  This comprises work towards [#1300] (data
auditing).

- Rename and enhancements of TimeStampedAbstractEntity to
AuditedAbstractEntity, which helps subclasses maintain a generic change log
(list) of "entity update" admin activities.  This will allow us to track
changes to any fields in an entity, in a generic manner, without having to add
specific relationships from the updated entity to a field-specific
AdministrativeActivity (this would be too cumbersome).  All
AuditedAbstractEntity classes now have a related "*_entity_updates" table in
the db schema.

- Added AdministrativeActivityType.ENTITY_UPDATE enum value.

- Removed AdministrativeActivity.{approvedBy,dateApproved} properties, since
these were essentially redundant with the performedBy and dateOfActivity
properties.  In cases where these approval fields were previously used, they
are now stored in performedBy and dateOfActivity properties, and the latter
fields are now stored in the AuditedAbstractEntity.{dateCreated,recordedBy}
properties.  This simplifies the semantics of AdministrativeActivity.

- One of the AdministrativeActivity constructors now has a recordedBy param,
to be used when the admin recording the activity is different than the admin
that performed the activity (e.g., for approvals, decisions, etc., rather than
direct data edits)

- Removed AdministratorUser.activitiesApproved, which was never used and would
now have been redundant with ScreensaverUser.activitiesPerformed.

- Added ScreenResultsDAO arg to ScreenResultParser, which was broken due to
this omission.

- Updated/fixed many unit tests

[#2104] fix loading failures for screen results (iccbl_support)
addressing
[#2119] Optimize ScreenResultParser for scalability with large import sets
- 1. Do not cache the ResultValues on the ResultValueType objects during parsing

minor refactorings and TODO notes

minor edits to small molecule user roles meta data

- fixed older WDAPTest unit tests that broke after introducing to new AbstractTransactionalSpringContextTests superclass
- removed unnecessary persistence-related code, where flushing & loading of entities is no longer needed

fixed older WDAPTest unit tests that broke after introducing to new AbstractTransactionalSpringContextTests superclass

fixed recently broken WDAP unit test

WDAP.testOnlyMutualPositivesAreShared() now passes

WDAP.testPairOfScreensWithMutualWellsButDisjointPositivesNotConsideredMutual() now passes

Fix bug whereby newly created screen could not be saved (PropertyValueException: not-null property references a null or transient value: edu.harvard.med.screensaver.model.screens.Screen.leadScreener)

Will work on optimization work for the ScreenResultImporter CLI.

- Implemented data access policy for mutual positives (i.e., positive ResultValues with overlapping wells) in WDAP
- Updates, fixes, refactorings to WDAPTest

WDAP.testSmallMoleculeUserLevel1ScreenPermissions() now passes

- Resolved [#2099] (error when adding second attached file of duplicate type)
- Resolved [#1709] (screen viewer attached file upload duplicate error message)
- Added GenericEntityDAO.merge() operation, which is needed to handle case where we need to reattach an entity network that contains multiple object instances of the same entity (in this case, AttachedFileType).
- Used GEDAO.merge() in {User,ScreenDetail}Viewer.save() methods, allowing us to fold together the 2 cases of saving a new entity versus updating an existing entity.
- AttachedFiles backing bean now uses UISelectOneBean instead of UISelectOneEntityBean, since the latter caused NonDupObjectExceptions in our save() methods (not directly related to the merge() call).
- Removed unique constraint on AttachedFile, since it was invalid to begin with, and code can now handle duplicate names for attached files for a given entity

fixed bug whereby attached files with same file name were colliding with one another, causing only one of the attached files to be displayed in the UI, even though all files were being managed by the parent entity

fixed bug whereby attached files with same file name were colliding with one another, causing only one of the attached files to be displayed in the UI, even though all files were being managed by the parent entity 

- Added update() to WebDataAccessPolicy, to allow permissions to be recalculated when data changes
- WDAPTest calls WDAP.update() when changing "my" user's association with her screens
- testSmallMoleculeUserLevel3ScreenPermissions() now passes

Renamed column z_ordinal to zdepth_ordinal, to fix problem that hibernate looked for a column zordinal in the database.

merge -r3520:3523 http://forge.abcd.harvard.edu/svn/screensaver/branches/nki/rev_3449 /home/cor/ws_scr_nki_trunk/scr_nki_trunk
(Bug fixing concerning
1) "N" forwarded to cellHTS2 in stead of "NS"
2) in case of "N" cellHTS2 takes both "N" en "NS" wells for calculating median of the negative controles/)

(1) "NS" replaced by "S" for annotating Shared Negative Controles. CellHTS2 took both wells in case of "N" negative normalization.
(2) When selecting "NS" in the webbroser, "N" was forwarded to cellHTS2 in stead of "NS". With change (1) now "S" is forwarded.
(3) Testing Negative "N" is now done with data also including a well with "S" annotation

corrected release date

Created branch to solve bugs: 
1) "N" forwarded to cellHTS2 in stead of "NS"
2) in case of "N" cellHTS2 takes both "N" en "NS" wells for calculating median of the negative controles/

- Removed "Administrative" data table column group.  Admin users would often
fail to notice that additional columns existed within the expandable subtree
in the UI.  Columns that were administrative continue to be restricted to
non-admin users.  TableColumn now maintains an explicit isAdminstrative
property, rather than inspecting the column group name.

- Factored out HasVocabulary interface from VocabularyColumn, and
RelatedEntityColumn now implements this interface.  This allows
RelatedEntityColumn to wrap VocabularyColumns.

- RelatedEntityColumn now correctly sets isAdminstrative property

Update news, change log for 1.8.1

- Resolved [#2098] (add screen summary columns to lab activities browser)
- Refactored ReagentEntityColumn into more general-purpose RelatedEntityColumn, which is now also used by LabActivitiesSearchResult to reuse screen columns defined in ScreenSearchResults.

Some minor changes to the screensaver install notes.

changed entity-type-association for an attached file type

updated version number, release notes, and change log for v1.8.1

cleaned import

add log output to report when database password has been acquired from .pgpass file

updated debug log settings

- Resolved [#2043] (do not allow a new screener to be entered w/o a lab head)
- Only labHeadAdmins can set a screener's lab to null.  UserAdmins that are non-labHeadAdmins must specify a lab on new user creation, and cannot edit lab on existing users.

Resolved [#2057] (Add an "About Screensaver" link).  Made "Screensaver" text in footer a link to the open source project home page.

- Resolved [#1674] (link in users browser broken for user name containing apostrophe)
- Removed <f:param> for linked cell values in in searchResult.xhtml, as we no longer use this param to communicate the current entity to the backing bean.  This fixes all problems occurring due to lack of escaped text in the param value string.
- Removed text-escaping hacks in AbstractBackingBean

- Resolved [#2032] (add "user role" column to UserSearchResults)
- Made UserSearchResults "Login ID" column now hidden by default
- Renamed ScreenerSearchResults "Lab Affiliation" column to "Lab Affiliation Category", moved after "Lab Name", and now hidden by default
- Added/updated some more entity relationship definitions

- Update for [#2024] (Add capability for PMCID information): changed internal storage of PMCID from text to integer, with UI converter that adds/removes "PMC" prefix as necessary for display/input.

Resolved [#2024] (Add capability for PMCID information)

- Resolved [#2070] (new Attached File Type values)
- Converted AttachedFileType from enum into an entity, to allow values to be modified via database updates
- AttachedFileType has {Screen,User}AttachedFileType subclasses, for type safety, since values are only applicable for eithers screens or screening room users
- Added a few more static RelationshipPath definitions to entity classes

Branch for development of 1.8.1 release, and subsequent maintenance, as necessary

Renaming library-mgmt-rework branch to dev-1.8.0, which is now a maintenance branch if further fixes are needed prior to release of 1.8.1

- WebDataAccessPolicyTest no longer uses CurrentScreensaverUser to set the current user, and instead instantiates new WDAP instances when changing the current user.  This means we no longer use AbstractEntity.isRestricted(), and must call the appropriate DAP.visit() method instead.  This is a more direct test of WDAP, but more importantly, allows us to change the current user within a given unit test.
- AbstractEntity.isRestricted() now throws exception if no data access policy is set (for safety)
- CurrentScreensaverUser now longer allows the user to be changed, once set.  Although this should never actually occur, this is a safety measure that is now necessary since WDAP has state that is specific to the current user, so changing the user would create security problems.
- Fixed more unit tests in WDAPTest

- Removed Screen.shareable and ScreenResult.shareable properties, as they have been subsumed by Screen.dataSharingLevel.
- Added migration script for these property changes.

added assay_well.version field to migration script

- Removed ScreenResult.shareable property, since this has been replaced by Screen.dataSharingLevel
- Updated WebDataAccessPolicyTest.rnaiScreenPermissions unit test
- Fixed WebDataAccessPolicyTest.mutualUserViewingPermissions unit test

WebDataAccessPolicyTest:
  - fixed  unit tests: testRnaiScreenPermissions(), testChemDiv6Restriction()
  - some data initialization refactoring

fixes for WebDataAccessPolicyTest, to reset testing context for each unit test (needed dataAccessPolicy to be "prototype" scope, and bug in tearDown method usage)

merge -r3464:HEAD http://forge.abcd.harvard.edu/svn/screensaver/branches/iccbl/library-mgmt-rework

- [#1296] WebDataAccessPolicy now calculates and caches screen access
permissions (lazily) for all data sharing levels.

- Added ScreenResult.assayWells relationship.  ResultValue constructors and
ResultValueType.createResultValue() factory methods now take AssayWell instead
of Well.  ScreenResult.createAssayWell() used to instantiate new AssayWells.
Updated affected code in ScreenResultParser, and CellHTS2 (both untested)

- Added ScreenResultsDAO.findAssayWell() operation, but currently unimplemented in
ScreenResultsDAOImpl

- AssayWell is now a mutable entity, to support changing of "positive" property

- Screen default dataSharingLevel set to "private".  Fixed Hibernate mapping
for this property.

- Enhancements, changes to WebDataAccessPolicyTest.  Unit test
"testSmallMoleculeUserLevel2ScreenPermissions" passes!

- WebDataAccessPolicyTest now extends AbstractTransactionalSpringContextTests,
which simplifies transaction management and allows entire tests to be run
within the same transaction.

- Work for [#2067] (create AssayWell entity).  Implementation sufficient for usage by WebDataAccessPolicy; to complete, still need to implement AssayWell.resultValues collection, replacing Well.resultValues, and replace screenResult.wells with screenResult.assayWells.
- moved ScreenDataSharingLevel from test/ to src/

- Added Screen.dataSharingLevel property and ScreenDataSharingLevel enum type
- Implemented screen data setup for screen sharing level unit tests

Refactoring of unit tests for screen sharing permissions among users and screens, consolidating user level 1 and 2 logic, and simplifying user level 3 logic.  Added logic for testing others published screens always visible.

- Added small molecule screener roles for mutual data sharing levels
- Draft implementation of unit tests for screen sharing permissions among users and screens of various data sharing levels

added "small molecule user agreement account disabled" notification message to login-error.xhtml, copying from login.xhtml, and made these messages standout in beautiful fuschia.

stubbed out tests for Data Access Level requirements (some of these tests will be moved to other test classes, as appropriate, when implemented)

Solved bug that it was looking for column zordinal in stead of z_ordinal, based on default. In the screen it is not yet shown "Unknown property 'zOrdinal'. Maybe rename column to depth_ordinal so it falls into the defaults of hibernate.

converted some log output to debug level

fixed typo

Added message about disabled accounts for users that did not sign the small molecule user agreement.

Added columns time_point_ordinal and z_ordinal for storing information high content imaging.

Resolved [#2037] (smiles and inchi should not be required by small molecule library contents parser)

Resolved [#2036] (SD file record header is too strict)

- Fixed circular dependency Spring injection problem with WebDataAccessPolicy and GenericEntityDAO beans.  Made DataAccessPolicyInjectorPostLoadEventListener a BeanFactoryAware bean, so that it could lazily find the dao, which could not be injected due to the circular dependency.
- Removed hack in WebDataAccessPolicy that obtained dao from the CurrentScreensaverUser object.  It's true that we traded a hack for a hack, but it's a less intrusive one.

updates to the PP Components: implemnent the "fail_fast" flag for records that have critical errors (no smiles...)

Branch for development of "Data Sharing Levels" feature

Removed compareTo() method from SmallMoleculeReagent, which was a carryover from the earlier Compound class, but which no longer makes sense (we used to need this to sort multiple compounds within a single well, but wells can only have one SmallMoleculeReagent now).

- cmd-line scripts now use the java executable that is found in the user's path
- removed obsolete scripts

Removed e.h.m.s.model.meta.{CollectionOfValues,Property}.  These classes have been subsumed by RelationshipPath and PropertyPath, resulting in a much simpler design and is easier to comprehend.

Removing plate-size-rework branch, since its development is complete and has been merged to /trunk

Aborting further development on the data-table-refactor, which never got very far in any case, and would now be too cumbersome to merge into the current version of Screensaver (v1.8.0).  Note that the only real change was the factoring out of criteria-related methods from TableColumn into a Criteria class.

fixed unit test

ignore src/edu/mit/ directory tree, which is generated by chembank-wsimport Ant target

updated changelog, release notes for v1.8.0

removed erroneously committed file

removed unused data members

Removed remnants of the transient Well.iccbNumber property, which has been replaced by Well.facilityId

Resolved [#2015] ("Browse Screens" broken for marcusAdmin users).  Required a hack to fetch Screen.fundingSupports from within DataAccessPolicy, since this relationship is not guaranteed to be initialized at time of execution (depends upon the calling code).

* merged the added insert in the schema_history for added column sql script, to the NKI branch.
merge --depth=infinity -r3447:3448 http://forge.abcd.harvard.edu/svn/screensaver/trunk /home/cor/ws_scr_nki_trunk/scr_nki_trunk
    --- Merging r3447 through r3448 into /home/cor/ws_scr_nki_trunk/scr_nki_trunk
    U  /home/cor/ws_scr_nki_trunk/scr_nki_trunk/database/manual_schema_updates/3243.sql

 [#2010] build fails if image/structure-images does not exist

[#1967] restrict access to structure image URLs
implemented using servlet filter
- more informative log message

removed obsolete cmd-line app

Added "miRNA" silencing reagent type, since we have "miRNA" RNAi library types

updated exception type and comment in CPRAllocator

[#1967] restrict access to structure image URLs
implemented using servlet filter
- updated the url-pattern to be more selective

- Removed constraint on CPR.addLabCherryPick() that prevented LCPs for the
same well from being added; this was wrong since it is valid for LCPs to exist
for the same well if they are on separate assay plate "attempts".  This fixes
[#2012] (CPR Viewer "Record Failed Assay Plates" command broken)

- Reinstated commented-out constraint in LCP constructor to prevent LCP
instantiation if source well has no reagent.  

- CPRViewer and CPRCherryPicksAdder now allow a SCP to be added even if it
would have no LCP(s), due to LCP source wells having no reagent. Also, 
duplicate LCP source wells are allowed.  Together, this allows
us to faithfully record a screener's *requested* cherry picks, and allows
admins to deal with this problem at a process level, rather than preventing
the SCPs from being added at all.  

- CPRViewer shows a CP admin warning if SCP has no LCP(s), which can occur if
the potential LCP(s) do not have reagents.

- CPRAllocator defensively checks whether a LCP source well has a reagent, to
handle potential legacy data where LCPs exist w/o having a reagent.  For
[#2012].

resolved [#2011] (CPR viewer "download" command is broken for CP w/o reagent)

force use of microliter units on output of volume in CPRPlateMapFilesBuilder

fix double logout "feature" (i.e. send redirect only when appropriate)

fixed logic error and NPE in LibraryAndPlatesUsed

corrected an old schema_history.screensaver_revision insert in 1949.sql, which has the wrong revision number (required for proper reporting by schema_updates_required.pl)

added update to schema_history table in 3431.sql

[#1967] restrict access to structure image URLs
implemented using servlet filter

2nd attempt at renaming lib-upgrade.sql to current revision number (3431.sql)

renamed lib-upgrade.sql to 3250.sql

fix non-performant delete of the image directory... old task uses file scanner to exclude structure-images, now using move task to temporarily move this dir out of the way.

Renamed script for adding owner column towards naming convention of revision.

WorkbookLibraryContentsParser now reports error locations as workbook cell locations

Resolved [#840] (make the availability of the cellHTS2 feature configurable)

Followed advice Andrew solving the bug "After adding or editing a library, open the editing window for another library, will give as owner the person in the first library.  It seems that the LibraryDetailViewer still has as the UISelectOneBean from the previous one. Because it is not null, it will not recreate it."

Advice Andrew:

The general pattern we've followed is to reset a Viewer bean's instance members when a setEntity(Entity e) method is called.  So in the case of LibraryDetailViewer, I would simply set the 'owner' data member to null in setLibrary(Library library), forcing the UISelectOneBean to be reinitialized (lazily) the next time getOwner() is called.

This "manual reset" pattern is neceessary because the viewer beans are instantiated once per user session, and so are reused between viewing of different entities.  (Eventually, we hope to simplify matters by using a "conversation scope" for JSF viewer beans, so that they're newly instantiated when appropriate, but that's a while off.)

- removed "unused field" log warning output in SDRecordParser, since it was too verbose
- no longer log parse exceptions, since they will be reported later anyway

fixed max file upload size problem, where it was not accepting files over 100MB

Added "Reagent ID" and "Sequence" fields to "Duplex Wells" table in Well Viewer

fixed unit test

fixed bug preventing viewing of a screen result's heatmaps for other than the first plate

Remove obsolete upgrader cmd-line apps for [#1362] (remove network I/O from library contents loaders)

fixed error with erroneously trying to determine numericalness of RVT during incremental screen result import

- Removed "Download Error-Annotated Workbook" feature, since we're moving away from workbook-based I/O and the feature was not working properly anymore.
- Fixed unit tests

fixed unit test

- Fixed NPE with Reagent.vendorId, which now always returns a non-null RVI (null object pattern).
- Removed null checks where individual fixes were made to handle this null case (no longer needed).

removed redundant method call

added debug log output for profiling support

- Performance improvements to LibraryContentsVersionManager.releaseLibraryContentsVersion().  Performs operation in constant memory space by flushing Hibernate session cache for each library plate processed.
- Made LibraryContentsVersion.release() take an 'updateInMemory' arg, to allow client code to decide whether the operation can be performed in memory for entire library.  If not, does not update wells' latestReleasedReagent, and assumes LCVM.releaseLibraryContentsVersion() is calling this code.

[#1967] restrict access to structure image URLs

updates for the pilot build, added "start-only", changed deploy task so the "structure-image" directory isn't erased.

Added selection field owner in case of adding and/or editing library. Or just for displaying.

fixed WorkLibraryContentsParsers to allow them to handle worksheets with trailing empty (but defined) lines; this was preventing multiple worksheets from being parsed

Fixed worksheet iteration bug that was preventing multiple worksheets from being parsed

Fixed parsing errors when encountering empty/null values for set-based gene fields.  We now instantiate a new {vendor,facility} gene, even if all fields are empty, for simplicity.

- Redesigned WorkbookLibraryContentsParser to reduce memory consumption by
converting workbook data to a text file prior to parsing.  This allows the
workbook to be closed, allowing of the associated memory to be freed and 
made available for the parsing operation, which can be considerable
(consider that there is a ~15x file-size-to-RAM memory usage multiplier!)

- ParseError is now a concrete class, which eliminates need for FileParseError
subclass, which is now deleted.

- RNAiLibrayContentsParser and NaturalProductsLibraryContentsParser have been
reworked to operate on simple text file string values, rather than
workbook2.Cell objects.  Refactored Column enum into a class hierarchy in
io.parseutil package, to allow for code reuse and for subtyping in order to
handle particular value types.

- VocabularyUserType.getTermForValueCaseInsensitive() now handles null value
by returning null, and throws exception if string value is invalid (simplifies
usage in parsing code).

[#1963] add "molecular weight" to small molecule reagent

[#1962] change library import file spec list delimiter to semicolon
[#1963] add "molecular weight" to small molecule reagent

[#1962] change library import file spec list delimiter to semicolon
[#1963] add "molecular weight" to small molecule reagent

[#1963] add "molecular weight" to small molecule reagent

- Repurposed LibraryContentsVersionReleaser into more general LibraryContentsVersionManager, which now is used to create and release LCVs.
- WorkbookLibraryContentsParser converts RuntimeExceptions during parsing into ParseExceptions so that they are not fatal.

upgraded postgresql jdbc driver

corrected NPE in RNAiLibraryContentsParser that can occur with empty multi-valued fields

correct NPE in WellSearchResults for species column

More parsing-related error handling code simplifications.

- For [#1177], [#1581]

- Refactored LibraryContentsParsers to function as one-record-at-a-time
parsers, via parseNextRecord(), which replaces the all-at-once parse() method.
This allows LibraryContentsLoader to handle common data persistence
operations, in particular, optimizations that are universally needed.

- This refactoring allowed all parsing-related error handling code to be
simplified.  All parsers now just throw errors, and no longer manage their own
error sets.  LibraryContentsLoader now manages errors for all parsers.
Standardized parsing exceptions so that all parsers just use ParseException.

- Optimized library loading by flushing and clearing the Hibernate session
cache every LCL.FLUSH_BATCH_SIZE records, so that memory consumption is
constant (and not linear w.r.t. the size of the library).  (Unfortunately, it
seems that the JDBC driver consumes memory linearly, so we have not achieve
constant-space performance, but this as good as we can do).

- Optimized library loading by preventing a well's extant reagent set from
being loaded.  Only the new reagent should exist in memory now.

[#1962] change library import file spec list delimiter to semicolon

merge --depth=infinity -r3237:3388 http://forge.abcd.harvard.edu/svn/screensaver/trunk /home/cor/ws_scr_nki_trunk/scr_nki_trunk
    --- Merging r3237 through r3388 into /home/cor/ws_scr_nki_trunk/scr_nki_trunk
    U  /home/cor/ws_scr_nki_trunk/scr_nki_trunk/test/edu/harvard/med/screensaver/model/propertytesters/AbstractPropertyTester.java
    U  /home/cor/ws_scr_nki_trunk/scr_nki_trunk/test/edu/harvard/med/screensaver/model/AbstractEntityInstanceTest.java
    Merge complete.
    ===== File Statistics: =====
    Updated: 2

[#1362] remove network I/O from library contents loaders

updated small molecule library contents test file to meet file spec for multi-valued fields

correct NPE in WellSearchResults due to missing reagent vendor identifiers (in legacy ICCB-L data)

resolved [#1990] (well search results workbook export should use multiple worksheets when needed)

reverted changes for [#1979], since this ticket is no a requirement

changed UI labels for library contents version from "Version" to "Library Contents Version"

[#1962] change library import file spec list delimiter to semicolon

Changed WellSearchResults "Duplex Wells" column to specify a PropertyPath, rather than a RelationshipPath, to ensure that the VirtualPagingDataModel is always used.

Added alternate versions of Well.create*Reagent() methods that take a 'updateReagentRelationship' flag argument, which allows memory-intensive operations (such as library loaders) to operate more memory-efficiently.

fix for structure image base URL configuration

- structure image base URL, needed by PlateWellStructureImageProvider, is now configurable via screensaver.properties
- Moved OrchestraPropertyPlaceHolderConfigurer to e.h.m.iccbl package tree, since it is ICCBL-specific and renamed to more generic PropertyPlaceholderConfigurerWithDatabaseConfiguration
- refactored spring-context*.xml files: moved OrchestraPropertyPlaceHolderConfigurer, which makes screensaver.properties available to Spring beans, into spring-context-config.xml
- cleaned up screensaver.properties* and log4j.properties.debug files

Resolved [#1195] (show duplexes in {reagent,well} {viewer,browser})

in lib-upgrade.sql, uncommented sql that set wells' latest released regeant, and moved to end of script, so as to consider all inserted reagent tuples

[#1962] change library import file spec list delimiter to semicolon

check for null return on standardized smiles

[#1979] Make the Activity "Performed By" column on the Screen page an Admin only column

[#1362] remove network I/O from library contents loaders

[#1362] remove network I/O from library contents loaders

[#1362] remove network I/O from library contents loaders

removing src/edu/mit/... directories from source repository, since classes here are now built on-demand instead

- Removed the recently-added Silencing Reagent Type" column from WellSearchResults, due to problems (column needs getVocabularySelections() method)

- Fixed NPE for facility gene info in WellSearchResults, discovered after importing an RNAi library anew, since genes can be left null
- Added "Silencing Reagent Type" to WellSearchResults, to facilitate export of import-able library contents files

removed JDK 6 API calls

removed maxlength file upload limit, although this still doesn't appear to allow large file to be uploaded 

cleaned import

cleaned import

scalability improvements to WellsSdfDataExporter

fixed typo in lib-upgrade.sql

removed obsolete, required cmd-line arg from LibraryCreator, no longer needed now that we removed library contents loading ability from LibraryCreator

fixed administrator lookup bug in io.libraries.LibraryContentsLoader

Resolved [#1950] (error when specifying funding support on new screen)

This can be generated, does not need to be checked in:
[#1362] remove network I/O from library contents loaders

- Screen.setKeywords() now public, to fix UI bug with "Add Screen" command (the UI property uses the setter).
- Relaxed EntityPropertyTest's test of "no setter for CollectionOfValues", since the above change causes this test to fail, and it's not clear what the real value of this constraint is.

[#1362] remove network I/O from library contents loaders

library for the Pubchem Component interface

[#1362] remove network I/O from library contents loaders

added unique constraint to LibraryContentsVersion, to prevent duplicate version numbers within the same library

Updated to lib-upgrade.sql db migration script to address a ICCB-L legacy data anomaly.

Updated to lib-upgrade.sql db migration script to address pending schema changes and ICCB-L legacy data migration issues

Removed null constraint on SmallMoleculeReagent.{smiles,inchi} properties, since ICCB-L legacy data does not conform (some SM experimental wells have no available compound information).

corrected schema column name

Fixed model unit testing bug, which caused testing of @ManyToMany, unidirectional relationships to fail

While attempting to debug failed unit test for LibraryCreator, had the realization that the LibraryCreator service and cmd-line app classes could be drastically simplified by removing their contents loading capability.  Since we have a separate LibraryContentsLoader service and cmd-line app that can be used in series with LibraryCreator, this simplification is desirable.  Unit tests were simplified as well, and passed after this refactoring, w/o additional debugging work.

added missing FundingSupportTest

fixed unit test

Fixed data access policy for Marcus admins not being able to view Marcus screens (identified via failed unit test)

Merged v1.7.5 from branches/iccbl/trunk into this library-mgmt-rework branch:
  merge -r3078:HEAD http://forge.abcd.harvard.edu/svn/screensaver/branches/iccbl/trunk

Resolved [#1230] (test library data should include empty wells)

library contents import failure no longer commits to database

minor fixes/refactoring of constraints in RNAiLibraryContentsParser

minor fixes/refactoring of constraints in RNAiLibraryContentsParser

lib-upgrade.sql database migration now adds LCV release activities for first contents version of each library 

- ui.LibraryContentsImporter.importLibraryContents() is no longer
@Transactional, to ensure that parsing exceptions cause transaction to be
rolled back; previously failed parsing would still commit the new library
contents version

- RNAiContentsParser:
  - now catches all exceptions (including domain model exceptions, and not just ParseExceptions) and reports as parse errors
  - requires both vendor name and identifier to be specified together if reagent exists for well; can no longer have an "unused" vendor name

- SmallMoleculeLibraryContentsParser now only adds a reagent for experimental wells

- LibraryContentsLoader.loadLibraryContents() explicitly requires an
AdministratorUser arg, instead of more general ScreensaverUser

WellSdfWriter now outputs SD records with field names matching the import file specification

Added libray contents version number to SDF exporter

fixed transactional problem with WellViewer.viewWell() methods by using a WellViewer proxy reference

- WellsSdfDataExporter allows a particular LibraryContentsVersion to be specified for the exported wells's reagent data.
- WellSearchResults sets the appropriate LCV in WellsSdfDataExporter before invoking the export operation
- Removed WellSdfDataExporter (the singleton exporter), which was merely a convenience class

Fix to allow Well Viewer to display reagent vendor identifier for a particular library contents version

Well Viewer now can display reagent info for a particular library contents version, rather than just the latest released contents version.  This allows Well Viewer to show the appropriate version of the reagent when admin are viewing Well Search Results for a particular library contents version.

fixed LibrariesDAOImpl.findReagents() HQL to return just the Reagent entities

- added non-null constraints to RVI, since db queries depend upon this
- enhanced RequiredPropertyException

changed Small Molecule import file specification field "Facility ID" to "Facility Reagent ID", for consistency with "Vendor Reagent ID" field

fixed renderning of "Download SD File" command button in Well Viewer

fix to SDRecord Parser to import molfile end marker, and handle case where it might be missing

improved null value handling in WellSdfWriter

robustness, error handling improvements to SDRecord Parser

- Changed import file specification for Small Molecule libraries: "pubchem_id" is now "pubchem_cid"
- Added output fields to WellSdfWriter to match the import file specification for Small Molecule libraries
- Validate required fields on import of Small Molecule libraries
- Fixed import of molfile data, to preserve whitespace and newlines
- Removed obsolete fields from SDRecord
- Added test of molecular formula and mass import to unit tests

- Fixed WellSearchResults, which was erroneously finding and showing reagents
from library contents version other than the "latest released" version

- Added Well.latestReleasedReagent, to allow required HQL queries to be
generated. It is not otherwise possible to filter through only the latest
contents version reagent when specifying left joins from well to reagent.
This is more performant on queries, in any case, since we no longer need to
access library.latestReleasedContentVersion.

- Updated ReagentEntityColumn and subclasses to only eager fetch the path
needed for the reagent: usually this is Well.reagent, but when showing
non-released LCV it will be Well.reagents with a restriction on reagent's LCV.

- Library.setLatestReleasedContentsVersion() now updates each well's
latestReleasedReagent

- Added Well.getPendingReagent() to get the reagent that next, non-released;
useful when importing new library contents versions.

- Renamed Well.getReagent() (previously a derived property) to
Well.getLatestReleasedReagent().  

- Made Reagent non-proxy, since Well.getReagent() would return a proxy class,
which caused problems with its generified return value.

- LibrariesDAO.findReagents() no longer takes relationshipPaths param and now
takes latestReleasedVersionOnly flag.

- Added Reagent.version to Well Viewer and Well Search Results.  

- Added ReagentNameValueTable for version field.  Required some refactoring of
NameValueTable class hierarchy.

- Enhanced EntityDataFetcherTest to test when multiple library contents
versions exist.  Also fixed the testSortColumns() test, which wasn't actually
testing anything.  Updated testFilterOnColumns() to test filtering on related
entities, in particular Well.reagent, to catch the WellSearchResults problems
that were discussed above.

- Fixed Library Viewer "delete" command's "enabled" status.

- Error lines numbers reported by SmallMoleculeLibraryContentsParser.  
  - Added InvalidWellException.  Used by SmallMoleculeLibraryContentsParser.
  - ParseException manages line number

corrected top-level user message on parse failure

- Streamlined error/exception handling in SmallMoleculeLibraryContentsParser
- Stopped handling IOExceptions in parser, allowing this exception to propagate and be reported as system error
- SDRecordParser no longer implements Iterator, since iterator methods could not throw IOExceptions.  This also fixes the problem of hasNext() throwing parse exceptions, which were not being caught by calling code, since the try/catch is inside the iteration loop, and so errors in the first parsed record would not caught and ended up being reported as system errors in the web UI, rather than parse errors

fixed NPE manifesting in Well Viewer

- fixed incomplete parsding of molfile in SDRecordParser
- added tests for molfile, chembank IDto SmallMoleculeLibraryParser unit tests

Reworked SDRecordParser to have hasNext() and next() iterator methods consider full records in the determination of "hasNext" status.  This avoids problems with empty trailing lines and other potential end-of-file detritus.  Performed many refactoringsto make the parser implementation simpler.

corrected library importer error message

fixed WellSearchResults, allowing a Library with no contents version to be viewed.

- Resolved [#1954] (add Library.isPool field)
- Removed Library.ParseOptions, which was only used to track whether imported library contents should be treated as a pool or duplex library.  This is no longer necessary now that we have the explicit Library.isPool field.

- fixed unit LibraryContentsLoadTest.doTestCleanDataRNAi_withDuplex() unit test
- SilencingReagent.getDuplexWells() no longer returns null elements if duplex well exists, but well has no (released) reagent (not needed by above unit test fix)

removed cruft from ScreensaverServletFilter

- Fixes to CPRViewer and CPRCherryPicksAdder to reinstate "Add Screener Cherry Picks" feature, with pool-to-duplex deconvolution feature.
- lib-upgrade.sql creates pool-to-duplex mappings in database for initial library contents versions.

Implemented a new StructureImageProvider to serve images from local filesystem, using filename based upon Plate and Well.

- Scalability improvements for data table export, for both worksheet (XLS) and SDF formats.  uses 

- Resolved [#1209] (library contents viewing permissions)
  - Added isAllowedAccessToSilencingReagentSequence() to DataAccessPolicy, to restrict siRNA sequences (a hack, to be sure, but okay for now)
  - e.h.m.iccbl.screensaver.policy.WebDataAccessPolicy restricts user access to:
    - SmallMoleculeReagent entities that are in ChemDiv6 library, unless readEverythingAdmin
    - NaturalProductReagent entities that are in Marcus libraries, unless marcusAdmin or readEverythingAdmin
    - SilencingReagent.sequence properties, unless readEverythingAdmin
  - WellSdfWriter respects DataAccessPolicy for SmallMoleculeReagent restrictions
  - WellViewer and WellSearchResults show silencing reagent sequence field, when allowed
  - WellViewer and WellSearchResults respect DataAccessPolicy for SmallMoleculeReagent and SilencingReagent.sequence restrictions
- WellsSdfDataExporter partitions database fetches, for scalability

LibraryContentsImporter clears comments field on subsequent usage

Implemented "Loading comments" field to LibraryContentsLoader

removed old, unused UI in libraryViewerBody.xhtml 

Implemented "Browse" feature for Library Contents Versions, allowing administrators to view the contents of any contents version.

- Resolved [#1907] (library contents version "release")
- Added LibraryContentsVersionReleaser service
- Library Viewer updated to use LibraryContentsVersionReleaser
- Fixes to Library Viewer UI, to display LCV release info and disable command buttons as necessary
- Removed duplicated reagent check in RNAiLibraryContentsParser, since this should never occur for a given LibraryContentsVersion (and if it does, domain model constraints will catch it)
- Added OperationRestrictionException, to be used for authorization violations when invoking service methods.  UIControllerMethodExceptionHandlerAspect now explicitly handles this exception.
- Fixed Well libraryWellType mutability constraints, only enforcing logic if type is actually changed; updated LibrariesDAOTest to test this.
- Fixed LibraryContentsLoaderTest, needed to release the loaded LCV, so that test logic would work

- Resolved [#1908] (delete library contents version)
- Implemented LibrariesDAO.deleteLibraryContentsVersion()
- Well now allows libraryWellType to be reset back to 'undefined' in all cases, allowing well type to be reset if/when sole library contents version is deleted.
- Removed LibrariesDAO.findLibrariesOfType(), as it was unused

fixed hibernate mapping problem with Gene entity and database sequence name

- Fixes to LibraryContentsImporter, making it functional
- LibraryContentsImporter no longer releases the loaded contents version, as this is done at the discretion of an admin later on.

- Factored out LibraryContentsLoader from LibraryCreator, so that LibraryCreator is no longer implemented the contents loading function.
- Removed GEDAO param from LibraryParser interface, and introduced into constructors in implementing classes (GEDAO is a behind-the-scenes detail, which the parser interface should not know about); removed static methods from LibraryParser concrete classes
- Some code reformatting

- Folded library-specific importer web pages into a single libraryContentsImporter.xhtml page
- Resolved [#1610] (remove LibraryUploadThread)
- More work towards [#1907](library contents version "release") and [#1908] (delete library contents version) in LibraryViewer

- Work towards:
  - [#1765] (library contents versioning)
  - [#1907] (library contents version "release)
  - [#1908] (delete library contents version)
- Added "Contents Version" panel to LibraryViewer, with data table displaying LCVs, and with Browse, Delete, and Release buttons in each row (as of yet, non-functional)
  

- Work towards [#1907] (library contents version "release)
- Added AdministrativeActivityType.LIBRARY_CONTENTS_VERSION_RELEASE
- Added LibraryContentsVersion.releasedActivity relationship.  Entity now mutable, with version field.  Also converted to a TimeStampedAbstractEntity.
- LibraryContentsVersion.release() method is now the API method used to "release" a LCV; Library.setLatestReleasedLCV() now only has package access.

removed commented-out code

- reinstated "Show Failed Lab Cherry Picks" feature in CPRViewer, requiring a few different fixes
- genericized some variables

- Final work to resolve for [#1256] (display 'reagent source id' constituent parts): CPRV and CPRExporter now place vendor name and vendor ID in independent fields.
- Fixes to CPRSearchResults and CPRViewer data table columns, now using only statically-defined RelationshipPaths to correct errors

- Reinstated CherryPickRequestViewer
- Required refactoring of WellSearchResults.WellReagentEntityColumn class into a generic ReagentEntityColumn class, with {Well,LabCherryPick,ScreenerCherryPick}ReagentEntityColumn classes (that latter two being used by CherryPickRequestViewer).
- Added explicit constructor to RelationshipPath for 0-length path (just the source entity, without a hop)
- Added more static RelationshipPath definitions to model entity classes

- more clean up of code concerned with library types; explicit handling of Natural Product library types, where necessary
- Libraries Browser, Library Viewer, and Well Viewer now working again for Natural Product library types

- minor clean up of code concerned with library types; explicit handling of Natural Product library types, where necessary
- added Library.getReagentType() method

removed ReagentVendorIdentifier.vendorName non-null constraint, which is no longer necessary since RVIs are no longer used as primary/foreign keys (now using reagent_id surrogate key instead)

- Added NaturalProductReagent to DataAccessPolicy
- removed a TODO in WellNameValueTable, as there was in fact nothing to do

Added "Duplex Wells" table to Well Viewer

Reinstated Study Viewer's reagents data table, now as a well search results

UI formatting tweak

- Reinstated Well Viewer's "Other Wells With Reagent" data display.  
- Added collapsible panels to Well Viewer for "Other Wells..." and "Annotations" data sections
- Added Screen Type field to Well Viewer name/value table, and reordered fields to match Wells Browser order
- Added RelationshipPath var args to LibrariesDAO.findReagents(), to support eager fetching.

- Reinstated "Find Reagents" feature.  Shows results as Well Search Results, showing all wells containing matching reagent identifiers.
- Changed LibrariesDAO.findReagent() to findReagents(), returning a set of Reagents (one for each well with matching reagent identifier)
- Added "Library Screen Type" column to Well Search Results, and placed "Library Name" column after Plate and Well columns.

final javadoc edits to resolve [#1270] (rename 'compound' to 'small molecule')

Final work to resolve [#1225] (RelationshipPath improvements): renamed getFormattedPath() to getPath(), and deprecated it, since client code should use RelationshipPath API directly to inspect the nodes comprising the path.

Use only facility gene information in the UI, and not vendor gene information (which we maintain for auditing purposes, but isn't otherwise needed by users at this time).

added TODO comments in lib-upgrade.sql to handle cases where reagent IDs exist for a well without reagent information, and vice versa

- Reinstated 'Download SD File' feature in Well Viewer
- Removed "structure image notice" from Well Viewer, since structure image rendering problems are no longer a problem (no no longer dynamically generated)
- Fixed some LazyInitEx in WellViewer, and performed some refactorings
- Fixed WellsSdfDataExporter eager fetching relationships

fixed LazyInitEx in WellViewer

reinstated molecular mass and formula fields in Well Viewer, [#1338]

Removed ReagentViewer, CompoundViewer, and GeneViewer.  WellViewer is now the sole viewer for this combined information, since we no longer have canonical reagents, compounds, or genes.

- Removed JSFUnit support in web.xml, which broke after recent changes to build.xml, and which is not being used anymore anyway
- Removed defunct bean from spring-context-services.xml

minor fixes to address FindBugs warnings

numerous, minor fixes to address FindBugs warnings

fixed WellVolume.equals() bug (kudos to FindBugs)

fixed bug where test code coverage runs would include test files after initial run.

- Fixed high-priority warnings identified by FindBugs, some of which were valid defects
- Added FindBugs annotations.jar
- Added findbugs-filter.xml, used to ignore test classes when invoking FindBugs

fixed alignment problem of result value columns in CherryPickRequestExporter for small molecule screens

removed all uptodate checks to test-coverage-* targets, since Ant provides an utterly broken system for determining out-of-date dependencies, and which still does not work in some cases.

Removed unused/obsolete PubChem querying classes.

moved ICCB-L specific package under e.h.m.iccbl package

- removed unused ScreenResultExporter and related classes in e.h.m.s.io.screenresults (to modestly improve test code coverage percentage!)
- moved ExtantLibraryException to appropriate package

remove unused classes (to modestly improve test code coverage percentage!)

updates for 1.7.5 release (changelog, news, version number, example database)

Resolved [#1935] (add screen "pin transfer approval" administrative activity), after adding UI fields to Screen Viewer.

build.xml test-coverage-instrument target now only instruments src/ classes, and not also test/ classes, which should not be analyzed

Updated model for [#1935] (add screen "pin transfer approval" administrative activity)

- moved old Hibernate-related pedagogical classes to flotsam+jetsam
- removed test.txt

build.xml now properly invokes dependent tasks when running test-coverage-{instrument,run,report}.  You can now invoke test-coverage-report alone, and all dependent tasks will be executed only if out-of-date.

removed useless class

renamed "Well Type" to "Library Well Type" in UI, to clearly distinguish from "Assay Well Type"

Fixed 'test-coverage-run' build.xml target.

Added support for Cobertura test code coverage tool.  Added test-coverage-{instrument,run,report} target, but test-coverage-run is not yet working.  'test' target is also broken at this time, though not needed for test code coverage runs.

Resolved [#1890] (userActivity log session ID is no longer useful).  Full session ID is now logged, rather than just a suffix.

- Resolved [#1934] (screen funding supports managed in database)
- Resolved [#1898] (new funding sources)

updated docs

see previous commit

Added my install notes

Temporarily removed access for labhead owner to the library, because of error "no session" when trying to initiate proxy for lazy loading owner.getLab()

Added policy for visiting library.

Library parsing: support for Duplex Wells, also, parse options passing to the parser

Library parsing: support for Duplex Wells, also, parse options passing to the parser

convenience getRow()

addressing issues outlined in [#1917]
- implement LibraryParser.parse()
- Library creation command line switch: if loading library at the same time
- LibraryContentsVersion creation from the command line: new switch --ecommonsID
- also refactor LibraryContentLoader work to the LibraryCreator service.

Started implementing library content only visible for owner

refactor all of the parser/loader tests into one test

LibraryCreator and parser(s) refactoring; note, only NaturalProduct Test has to be fixed

merged to head version (=1.7.4?) canonical tree.
 
 svn merge http://forge.abcd.harvard.edu/svn/screensaver/branches/nki/trunk@3181 http://forge.abcd.harvard.edu/svn/screensaver/trunk@3237 /home/cor/ws_scr_nki_trunk/scr_nki_trunk > merge_canon_3181_nki_3237.log

simplified test class and method names

removed unused data member

- Fixed WellSearchResults, which broke for columns that are based upon Well.reagent relationship.
- Implemented a TableColumn wrapper, WellSearchResults.WellReagentTableColumn, which provides null and type safety checks, and adds the required Well.library.contentsVersions eager fetch relationship path
- This above required a major refactoring of the EntityColumn class hierarchy:
  - EntityColumn class removed, and its subclasses now directly extends their respective TableColumn subtypes (IntegerEntityColumn extends IntegerColumn).  This eliminates the parallel class hierarchy we had for entity-based and non-entity-based table columns (yes!).
  - Factored out HasFetchPaths interface from EntityColumn, as its interface methods were really the only extra behavior previously provided by EntityColumns.  *EntityColumn classes now all implement HasFetchPaths.  Code that used to rely upon EntityColumn now relies upon either TableColumn or HasFetchPaths types.  This reduces coupling to TableColumn types in code that was not really interested in TableColumns, per se, but only interested in fetch the data network needed to satiate those columns' data needs)
  - Added FetchPaths mixin implementation of HasFetchPaths, and is used by *EntityColumn classes, to eliminate redundant implementations
- RelationshipPath.to(RelationshipPath) method now (properly) appends the full path of the provided RelationshipPath, allowing two multi-hop paths to be concatenated.  Needed by WellReagentsTableColumn implementation.

update LibraryCreatorTest to handle RNAiLibraryFileContentsLoader

allow vendoridentifier to be null

- Make use of Library.experimentalWellCount in UI, replacing dynamic calculation of this value.
- Removed LibraryNameValueTable, which was no longer used.
- Updated lib-upgrade.sql to calculate Library.experimentalWellCount for existing libraries.

parser refactoring

parser refactoring

Resolved [#1292] (store pre-calculated library size)

UI fixes

Updated javadocs for GEDAO.findEntityById(), documenting its performance advantages for pre-fetched entities.

UI fixes, allowing WellSearchResults (and Browser) to work.

Merged latest from branches/iccbl/trunk, thus incorporating ICCB-L 1.7.4 production release:
merge -r3018:HEAD http://forge.abcd.harvard.edu/svn/screensaver/branches/iccbl/trunk

Fixed compilation errors discovered during web deploy
  - reverted to JDK 5 API usage, since we're not yet officially supporting JDK 6
  - removed dependency on test/ classes in MockDaoForScreenResultImporter

- Removed Library.compoundCount property, which was not used, and had been intended for maintaining values of legacy data (this value can otherwise be computed)
- Added some missing schema changes to lib-upgrade.sql

Added LibrariesDAO.findReagent() method, removed LibrariesDAO.findSilencingReagent() method.

removed schema migration scripts that are now integrated into lib-upgrade.sql

More work for [#1270] (rename 'compound' to 'small molecule')

More work for [#1270] (rename 'compound' to 'small molecule')

fix test for undefined wells

Fixed unit test for Well.libraryWellType, requiring custom test, due to conditional immutability constraint

relaxed well type mutability constraint, mostly to allow testing code to work as needed

fixed unit test assert, though unit test still fails

fixed unit tests in CherryPickRequestPlateMapperTest

test

fixed unit tests in WellsSdfDataExporterTest

fixed unit tests in CherryPickRequestDAOTest

fixed unit test

Added CPRCherryPicksAdderTest to respective PackageTestSuite

Added unit tests for CPRCherryPicksAdder, with some minor error-handling changes to the tested class

corrected spring-context-services.xml config error

finish functionality and tests for RNAiLibraryContentsParser et. al.

finish functionality and tests for RNAiLibraryContentsParser et. al.

only allow undefined wells to be set, to enforce immutability after versioned library creation.

default library wells as undefined not empty

- Initial rework of pool-to-duplex deconvolution logic, using new domain model structure.
- Removed LibraryPoolToDuplexMapper altogether.
- Moved CPRViewer cherry pick adding logic to CPRCherryPicksAdder service.

fixed unit tests

fixed unit test

fixed unit test in WellSearchResultsTest

- Fixed all but one unit tests in WellSearchResultsTest
- Converted RelationshipPath instantiations to use statically-defined model RelationshipPaths, adding many new ones, as needed 

toString method

fix internal access to uninitialized state variables on worksheet, workbook

- Fixed model unit tests
- 100% pass!

fix iterator index upper bound

fix internal access to uninitialized state variables on worksheet, workbook

AnnotationValue.reagent is now non-nullable; unit test now passes

- Fixed model unit tests
  - Added hack to EntityNetworkPersister, in order to persist WellVolumeAdjustments, which is a special-case (two mutually exclusive required to-one relationships)
  - Annotated entity relationships that must be tested with custom unit tests

- Implemented pool-to-duplex relationship in SilencingReagent, which is actually a pool-to-duplex-well relationship, to allow respective pool and duplex library contents versions to change without breaking the relationship.
- Renamed SilencingReagent.annotatedGene to facilityGene, for improved clarity and for consistency with RNAi import file format specification.
- Moved MakeDummyEntities from src/ to test/

added model unit testing annotations to allow some units tests to pass

change setters back to private on gene to preserve immutability constraint

Updated javadocs for model testing annotations

fixed bug in TestDataFactory.newInstanceUsingParentFactory(), fixing (at least one) unit test

fixed bug in TestDataFactory.newInstanceUsingParentFactory(), fixing (at least one) unit test

fixed unit test

MakeDummyEntities.makeDummyStudy() refactored to use TestDataFactory

- added missing cascade to AdministrativeActivity
- fixed AnnotationTypeTest unit test

For upload, added NS as an expected value for AssayWellType .
Also added example of it in the data for test on upload in case of channels.

updates for rnaiLibraryContentsParser refactoring; workbook2 refactoring; library rework

update for workbook2 refactoring

updates for rnaiLibraryContentsParser refactoring; workbook2 refactoring; library rework

workbook2 refactoring; library rework

updates for screenresultparser refactoring; workbook2 refactoring; library rework

put rnailibrary contents importer back in

updates for workbook2 refactorings; and for library rework model changes.

added hack to EntityNetworkPersister to handle persistence ordering problem (need to analyze ENP to determine if its a model or ENP deficiency); fixes some unit tests.

minor fixes

fixed some model unit tests by moving "ownership" of LabCherryPick from CherryPickRequest to ScreenerCherryPick; this eliminates the multiple-parenting problem, which was confusing the NetworkEntityPersister; more important, this change also implicitly enforces the model constraint that a SCP and its related LCPs must all share the same CPR.

fixed model unit test

fixes for model unit tests

ModelIntrospectionUtil.isNonNullableProperty returns false if no appropriate annotations can be found, instead of throwing exception; fixes (at least one) unit test

model fixes to pass unit tests

Implemented model unit tests for to-one entity relationships.

Implemented model unit tests for to-one entity relationships.

Fixed many custom entity factory methods, which were using the wrong newInstance*() method.

- reinstated running of all model unit tests, some of which were disabled temporarily during model unit test development efforts
- fixed testEqualsAndHashCode() test, which was not testing transient instance properly

model fixes to pass unit tests

fixes and refactorings to model unit testing framework to support to-many relationship testing

* Added NPI as option to plate normalization
* created RUnit and JUnit test for NPI plate normalization.

fix compilation error

- Model unit tests now handle the case where a to-many relationship between entity type A and B is required on the many-side, but where A is *not* the entity containing the create factory method.
- removed ContainedEntity 'alternate*' attributes from ContainedEntity, since model tests do not use these.  ContainedEntity is now only used to determine the parent entity where factory methods for the child entity can be found.

- Implemented model unit test for to-many entity relationships.
- Folded e.h.m.s.model.annotations.{Many,One}To{One,Many} into just ToOne and ToMany annotations.
- Some model fixes in response to failed unit tests

cleaned imports

EntityNetworkPersister had non-deterministic behavior due to use of non-ordered sets, and would fail intermittently.  Corrected by using lists instead, and now specific model unit tests fail consistently when attempting to persist entity network  So there are still some problems, but not yet sure if they ENP problems or domain model problems (ideally, if model problems, ENP should at least catch the problem and report it, but isn't).  Made a few minor refactorings as well for debugging efforts.

Added JUnit test for normalizePlatesNegControlsNegShared

fixed ScreensaverUser model unit test for roles, which required custom test in each concrete entity subclass

fixed Screen model unit test for billingItems, which required custom test

- Fixed model unit tests for Gene
- Fixed model unit tests for SmallMoleculeReagent, and fixes to SmallMoleculeReagent mappings.  
  - Added tests for related collections, since automated model tests cannot these immutable collections
  - Moved molfile unit test from WellTest to SMRTest
- Fixed compilation errors caused by AEIT.dataFactory visibility

- Performance optimization in model tests, by only truncating tables when a test is ready to use the database, since some tests decide to skip themselves and never hit the database.
- Other minor code cleanup and refactoring in model tests.

Added normalization by NEGATIVES to the 
Made preperations for the additional parameter 'negControls' in the cellHTS2 normalizePlates method. This parameter contains the annotation of the negative control to be used. In our case it will be used to select "NC" (negative control common) if present, otherwise just the N.

minor updates to model testing classes

minor entity corrections, to pass model unit tests

fixed unit test for ScreenResult, using new CollectionOfElements.hasNonconventionalMutation annotation

Added CollectionOfElements.hasNonconventionalMutation annotation attribute, to allow some collection of element property tests to be skipped by the automated model tests.

AbstractEntityInstanceTest subclasses now dynamically add individual tests for each bean property to be tested.  This provides better visibility of test coverage and of the total number of model tests that pass/fail (previously, a single failed property test for an entity would prevent remaining property tests from being run).  Implemented by moving property-related testing code from AEIT to new EntityPropertyTest class.  AEIT subclasses now call AEIT.buildTestSuite() to dynamically add the property tests.

added PlateSizeTest, to test PlateSize.getEdgeWellNames()

- Removed CollectionOfElements.initialCardinality annotation attribute
- This was only used by CPR.emptyAssayPlateWells property, and and now it no longer populates this by collection by default.  Instead, it is up client code to populate as necessary.  This is better, in any case, since the set of default empty assay plate wells is a matter of policy, and not a model-level concern.

Re-implemented model tests for CollectionOfElement properties.

* Adjusted configureDb so that it no langer put the well type annotation to lowercase.
* Added JUnit for normalization by "negatives"

fixes for Screen and LabHead model tests

fix for AnnotationValue model test, et. al.

fix for Publication model tests

fix for ChecklistItemEvent model tests

fix for ResultValue model test

fix for PlatesUsed model tests

fix for AnnotationTypeTest model test

fixes for SmallMoleculeReagent model tests

fixes for Well model tests

fixed model unit tests for Reagent

fixed some model unit tests by correcting EntityFactory classes for {SmallMolecule,NaturalProducts}Reagents

Changed Gene API for setting property values, in order to conform to design conventions expected by model unit tests.  Setters of this immutable class are now private, and instances can have its properties set via with*() builder methods, prior to being persisted.

fixes for model unit test

fix for model unit test

In TestDataFactory, selection of create*() factory method in parent bean now prefers method with most specific return type, to ensure the instantiated entity is of the requested type.

fixed some model unit tests by adding/updating EntityFactory classes

In the unit test, there was a fixed value used 

I have upgraded on yoda the cellHTS2 version from 2.4.1 to version 2.8.0.  There are not much changes. The main change is that when now running cellHTS2 the  generated pages of the reports are organized in tabs in stead of everything below each other. Pages generated before the upgrade will however still look the same.

The reason for upgrade is that on wooki and the future production machine we will run R 2.9.0. When installing cellHTS2 in R 2.9.0 we will get cellHTS2 2.8.0 by default. 

- Transposed how bean properties are tested.  Instead of performing each type of
property test on every bean property, we now iterate once over all bean
properties and apply the appropriate test, based upon its type (simple value,
to-one, to-many, collection of elements, etc.).  This eliminates considerable
amount of redundant logic.  It is also more transparent regarding the coverage
of a bean's properties (i.e., were all properties tested?).

- Removed PropertiesGetterAndSetterTest, since it was vacuous test.

- Removed CollectionPropertiesInitialCardinalityTester, since it was a
redundant test, having been tested in AbstractEntityInstanceTest already.

- Removed model.propertytesters packages and its member classes, since all
logic already in AbstractEntityInstanceTest

- Currently, entity relationship-based tests are commented out, but will be 
re-instated in upcoming work.  Simple value property-based tests are 
currently working.

- Refactored factory methods for test data

    - Moved all test data factory methods to TestDataFactor

    - Removed 'parentBean' and 'persist' args, since they were introducing
    considerable complexity, and their semantics and behavior were not
    well-defined.  Now calling code simply persists as necessary using the
    NetworkEntityPersister, and rewrote tests to not rely upon
    parentBean-based child entity instantiation.

    - Simplified newInstance*() methods, combining methods where possible, and
    eliminating redundant and confusuing logic.

- Refactored model introspection methods by moving into ModelIntrospectionUtil 
classes

- Merged concreteStandinMap into the more general entityFactoryMap, since the
latter provides a natural means for instantiating a concrete entity type for a
given abstract type.

- Removed special instantiation logic from entity test classes; will reinstate
with a more appropriate design in the future

update lib-upgrade.sql migration script for earlier change to make SilencingReagent.{annotated,vendor}Gene properties not-null

add columns iupac and cas_number to small_molecule_reagent

updates for field name changes

baseline unit test with test data

comments

add/update fields for the small molecule import:
iccb_number -> facilty_id

unit test fix and work for [#1909] (study annotations should annotate well, not reagent)

unit test fixes

unit test fix

unit test fixes

more fixes for model units tests, both to model classes and to unit test classes

- SilencingReagent.{vendor,annotated}Gene related entities are only created on demand now and are now immutable (thus, might be forever null if not set before first persist)
- Model fixes for unit tests
- AbstractEntityInstanceTest fixes and improvements, including new entity factories for Reagent subtypes and Gene.

unit test fixes related to [#1907] (library contents version "release")

Model updates for [#1907] (library contents version "release")

Fixed error in screenDetailsHelpBody.xhtml, which caused error when viewing content on Help page.

Branch for getting online documentation to work again.

fix the small molecule parser and test to work with library refactorings

- Changed RelationshipPath and PropertyPath API to eliminate string-based,
dot-separate path specifications, instead requiring paths to be built with
iterative calls.  This opens the door for compile-time checking of paths, by
using model-defined constants for each entity-to-entity, entity-to-property,
and entity-to-collection relationship.

- Created model.meta package, and moved meta-model-related classes there
(RelationshipPath, PropertyPath, RelatedProperty etc.)

- Added model.meta.{Property,CollectionOfValues} classes, used to specify
model meta info that can be compile-time checked

- Started adding RelationshipPath constants to the domain model classes, to
define entity-to-entity relationships as constants.

- Updated *SearchResults classes to use new {Relationship,Property}Path API.

- Fixed many unit tests, updating to work with recent Library domain model
changes and using {Relationship,Property}Paths and domain model relationship
constants where applicable.

Adjusted documentation

fix to lib-upgrade.sql

Added lib-upgrade.sql, which is a draft schema migration for the library rework branch data model.  Still some TODOs noted in the script, but largely complete and working on ICCB database.

AnnotationValue is now declared immutable and no longer needs 'version' property.

more work for [#1270] (rename 'compound' to 'small molecule')

More thorough annihilation of CAS and NSC numbers for [#1903]

Resolved [#1903] (remove CAS and NSC numbers from domain model)

fixes to domain model, as revealed by schema migration analysis

fixes to domain model, as revealed by schema migration analysis

corrected hibernate mapping for values collections in Gene and SmallMoleculeReagent

Fixed compilation errors reported by non-Eclipse (OpenJDK) Java compiler with generic method Well.getReagent().

- Added basic test to verify that library contents can be created using new domain model.  
- Made minor model fixes and improvements in response to problems found during test development and execution.

- Fixes to allow schema to be generated by Hibernate
- ReagentVendorIdentifier now requires vendor and reagent ID to be specified iva independent parameters (no longer parsing these out from a single string param)

updates to uml model

- Removed all remaining compilation errors by commenting out code and removing obsolete code.  Code is completely broken still, of course.
- Also reinstated POI/HSSF jars for now, since it turns out there are still non-library code dependencies on it. 
- Added "TODO library_rework" in most places where we need to go back and reinstate functionality.

For [#1310] (replace HSSF with jxl).  Delete the POI/HSSF jars.

For [#1310] (replace HSSF with jxl).  Delete the io.workbook package, which wraps HSSF library.  Dependencies on this library not yet fixed.

Added loess normalization

Added loess normalization

- Extensive rework of code throughout entire system that is affected by library domain model changes, correcting compilation errors only, and making no claims about correctness of code.  Ignored compilation errors in library parser/importer classes since these are about to be recoded altogether.  Just trying to clean up compilation in order to focus on forthcoming import/parser work, without compilation error distractions.
- Removed ReagentSearchResults class for [#1378] (cross-screen comparisons should be on reagent, not well)
- Removed classes that I don't think we'll ever use this again: BulkRNAiLibraryLoader, BioactivesSubmissionCompoundNameGrabber

added schema_updates_required.pl

The cfg/orchestra/tomcat directory is no longer needed, now that we have web/META-INF/context.xml

- Moved these files out of svn repository, as they are now located in ICCB-L's production environment, and really have no need to be versioned.  
- This also reduces the differences between the ICCB-L branch (/branches/iccbl/trunk) and the main trunk (/trunk), which minimizes branch merging complexity.

- fixed data access permissions violations in WebDataAccessPolicy, as detected by unit tests
- added unit test for marcusAdmin-related permissions in WebDataAccessPolicy

fixed unit test

added userChecklistItemsAdmin role to example user in initialize_example_users.sql

Updated example database with latest schema changes

updated change log for v1.7.4.

updated version number and change log for v1.7.4.

Added a feedback link to the error page.

Reinstated "back" button on error reporting page, which we had previously not been able to get working and had been disabled for so long, I can't bare to say.

Resolved [#1710] (improve handling of user login attempt w/o "screensaverUser" role)

Resolved [#1705] (add "Library" column to cpr viewer lcp and scp data tables)

Resolved [#1877] (add login help links).  Added "mailto" links, replacing existing, ineffectual link.

Resolved [#1876] (use consistent font)

Pulled up _isEditMode data member into AbstractEditableBackingBean from all of its subclasses.

Refactored property.xhtml for volume and concentration field types.

Resolved [#1862] (new cherry pick request fields)

- Resolved [#1869] (new admin role for editing user checklist items only).
- Added 'userChecklistItemsAdmin' role.
- Added ui.util.ChecklistItems backing bean, factoring out code from UserViewer (also groundwork for [#1295]).
- Added ChecklistItemsEntity interface, implemented by ScreeningRoomUser (also groundwork for [#1295]).
- Refactored JSF UI definition for Checklist Items section from userViewer.xhtml into checklistItems.xhtml.
- Fixed popup.xhtml and textColumn.xhtml, whose "contents" fragment had namespace collision with collapsiblePanel.xhtml and one another.

- Bug fix for [#1858] (support file attachments to user entity).  Fixes LazyInitEx when MarcusAdmin attempts to open LabActivity from Screen Viewer.

- Resolved [#1858] (support file attachments to user entity)
- Added ui.util.AttachedFiles backing bean, factoring out code from ScreenDetailViewer, and now using in UserViewer
- Refactored JSF UI definition for Attached Files section from screenDetailViewerBody.xhtml into attachedFiles.xhtml
- Moved AttachedFile from model.screens package up to model, since it is no longer screen-specific.

removed obsolete cfg/tomcat/conf directory tree

Attempt at fixing login.config file location problem in new ICCB-L production web environment.  java.security.auth.login.config property is now set in a deployable catalina.properties file.

- Resolved [#1856] (add 'Added to QPCR email list' user checklist item)
- Resolved [#1857] (marcus user agreement user checklist item changes)

- Resolved [#1859] (add 'marcusAdmin' role and rights)
- Added 'marcusAdmin' role
- WebDataAccessPolicy now restricts Screens, ScreenResults, CherryPickRequests, and LabActivities from marcusAdmins if the associated Screen does not have Marcus funding support.  This also indirectly restricts the sharing of non-marcus screen result data header columns in well/reagent search results (since RVTs assume ScreenResult permissions).
- UserViewer now obeys restricted status of Screens in screen tables.
- Screen.fundingSupports is now eagerly fetched, since we always need this to calculate whether a Screen is restricted
- Caveats: 
  - Screeners that are not related to Marcus screens can be viewed by marcusAdmins (but not their screens)
  - It is possible for a marcusAdmin with screensAdmin role to add a screen w/o the Marcus funding support, and thereby causing the screen to irretrievable.

Resolved [#1863] (remove molecular mass and formula fields).

- Resolved [#1363] (duplicate 'performedBy' for new screening activities).
- Refactored activity duplication logic into new ScreeningDuplicator service.

Changed JAAS realm configuration, moving screensaver.xml file to web/META-INF/context.xml.  This configuration works for both localhost Tomcat installations and on ICCB-L's orchestra production environment, and is required by the latter, which was the motivation for the change.

Resolved [#1673] (add "cherry pick request" column to activities browser).

- Resolved [#1774] (default volumes to uL in cprv)
- Resolved [#1861] (library screening activity viewer should default to nL units)

Resolved [#1368] (warn on leaving unsaved viewer page in edit mode).  Menu panel is hidden while in edit mode, thereby preventing accidental navigation to another page prior to saving.  Updated template.xhtml to accept a new 'hideMenu' parameter, used by editor pages.

Resolved [#1848] (migrate example database).  Applied missing schema migration scripts, dumped schema, and gzipped.

Resolved [#1693] (add "date performed" column to cpr browser).  Added column showing date of most recent CherryPickLiquidTransfer ("cherry pick plate activity").

Resolved [#1565] (link CherryPickPlateActivity to CPR)

Resolved [#1852] (add new role for "QPCR User")

Resolved [#1812] ("update user checklist items to record 3 tiers of small molecule user agreements")

Reverted changes from 3042 and 3041 in branches/iccbl/trunk, as these were changes unintended (they are changes made on separate project, and will be incorporated later).

- Improved how error messages are added to parsed workbooks. Workbook.addError() and Cell.addError() are now used exclusively for workbook-wide and cell-specific parse error messages, respectively.  Client code no longer has access to the workbook's ParseErrorManager.
- WorksheetColumnSpecification no longer throws WorkbookParseException, but rather is required to add a cell-specific parse error and return null, when necessary.  Eliminated WorkbookParseException
- Removed try/catch blocks from parsers, which were required by WorkbookColumnSpeficiation.parse() thrown exception

Added project-specific Workbook2 package (copied from Screensaver), allowing local improvements to be made

Resolved [#1757] (Add "COMS registration number" and "ICCB-L/NSRB COMS approval for this screen" fields for screens)

Small fix for migration script for [#1779] (Modify user checklist items related to user/datasharing agreements)

Resolved [#1779] (Modify user checklist items related to user/datasharing agreements) 

Resolved [#1767] (remove openbabel dependency from library contents loaders)

- SDRecordParser is now a simple parser.  
  - Implements Iterator for simplicity.  
  - Knows only about converting text from an input stream into a simple data structure.  
  - No longer knows about its parent loader, domain model, database access, error management, etc.  
  - All of these higher-level concerns moved up into SmallMoleculeLibraryContentsLoader.
  - LibraryContentsParser interface renamed to LibraryContentsLoader, to imply that data model is in fact updated (not just a parser)
- [#1270] (rename 'compound' to 'small molecule').  Renamed following to use "SmallMolecule" prefix:
  - io.libraries.smallmolecule.SDFileCompoundLibraryContentsParser
  - io.libraries.smallmolecule.CompoundLibraryContentsImporter 
  - ui.libraries.CompoundLibraryContentsImporter
- Deleted BulkCompoundLibraryLoader, since its purpose is beyond the scope of Screensaver; will add a cmd-line app for the SmallMoleculeLibraryContentsLoader
- PubChemIDs and ChemBankIDs now Integers instead of Strings
- [#1765] (library contents versioning)
  - Added Library.createContentsVersion()
  - Added AdministrativeActivityType.LIBRARY_CONTENTS_LOADING
  - Renamed 'versions' to 'contentsVersions' for clarity
- Resolved [#1254] (all LibraryContentsParser classes must respect startPlate/endPlate)
  - Simply removed the start/end plate ranges from the LibraryContentsLoader interface, as I suspect this optimization will no longer be needed now that library contents loaders are becoming simplified (no network I/O, calculations, etc.)

- Updated SDRecordParser to read in all required information directly from SD file (coded, not tested), addressing:
   - [#1362] (remove network I/O from library contents loaders)
   - [#1767] (remove openbabel dependency from library contents loaders)
   - [#1338] store molecular mass and molecular formula in the database, instead of recomputing them each page view
   - [#1351] sd file library contents importer should grab chembank ids
- [#1420] (SD file 'compound name' tag variants): now just "Chemical_Name" supported
- [#1306] (modify compound library contents importer to use hibernate cache): obsolete, now that Compound entity corefence no longer being attempted
- SDRecordParser is vastly simplified, as it becomes a simple text parser, without any network I/O, real-time molfile conversions, entity coreference, and primary compound determinations.

- resolved [#1772] (rename WellType to LibraryWellType)
- added libraryWellType to SDRecordData

[#1270] (rename 'compound' to 'small molecule'): renamed all 'compound' packages to 'smallmolecule'

commented out guts of CompoundPubchemCidListUpgrader, to eliminate compilation error, since this I/O service will be reincarnated as a stdout-only program (no db updating)

fixes for model refactorings [#1765], [#1766], [#1768]

[#1769] (add NaturalProductsReagent entity): updated uml model

Resolved [#1769] (add NaturalProductsReagent entity)

fixes for model refactorings [#1765], [#1766], [#1768]

fixes for model refactorings [#1765],  [#1766], [#1768], [#1586]

- Added Well.create{Silencing,SmallMolecule}Reagent() methods.  
- SilencingReagent automatically creates empty, related vendor and annotated gene entities, to ease instantation effort of SilencingReagent.
- model package is now free of compilation errors.

Initial, sweeping changes to model in library package.  Completely and utterly broken, and highly experimental.  This is an intermediate commit for safety reasons only.

- [#1270] (rename 'compound' to 'small molecule')
- [#1765] (library contents versioning)
   - Added LibraryContentsVersion entity, used by Library and Reagent
   - Well now has a map of Reagents, keyed on LibraryContentsVersion; logically, Well has a single Reagent at a particular point in time
- [#1768] (make Reagent base type of SmallMoleculeReagent (aka Compound) and SilencingReagent)
   - affected entities updated (only in the model packages so far)
- [#1766] (one-to-one well-compound relationship)
   - Reagent and SilencingReagent and SmallMolecule Reagent subclasses and child entities (Gene) now have containment relationship with Library/Well (no sharing of library contents or related entities between libraries)
   - these entities are now assigned a db ID and immutable
- [#1338] (store molecular mass and molecular formula in the database, instead of recomputing them each page view)
   - Added these as explicit properties to SmallMolecule
   - Updated importer for small molecule libraries
- [#1586] (wells should maintain both orig and latest entrez gene info)
   - SilencingReagent now has dual vendorGene and annotatedGene relationships to Gene entity
- [#1204] (model enhancements with Reagent entity)
   - SilencingReagent now has explicity pool and duplexes (self) relationships
- [#1767] remove openbabel dependency from library contents loaders
   - smiles, inchi moved to SilencingReagent
   - library loader now explicitly requires these values 
- Other:
   - Moved reagent-specific information from Well to appropriate class
      - smiles, molfile now in SmallMoleculeReagent
      - genbankAccessionNumber moved to Gene.genbankAccessionNumbers (exclusively)
      - gene now in SilencingReagent
   - removed legacy reagent information: oldEntrezGene{Ids,Symbols} (will handle via library contents versioning)
- added uml.zargo (nascent, for dev purposes only currently)

Removed CDK library, no longer used now that Compound molecular mass and formula are not calculated a runtime; for [#1338]

- Work towards [#1338] ("store molecular mass and molecular formula in the database, instead of recomputing them each page view")
- Added explicit molecular mass and formula fields to the Compound entity
- SDRecordParser partially updated to handle acquisition these fields from SDF file (to handle multiple mass/formula pairs)
- Added MolecularFormula{,Type} classes.

Resolved [#1291] ("remove model.derivatives package")

Resolved [#1494] ("review library entity requirements").  Removed unused library fields (by ICCB-L).  Renamed informaticsComments to just comments.

branch for library management rework (see tickets for iccbl::library_mgmt release)

Added option to select for scale people want to normalize. 

Resolved [#1733] (New User Checklist items: Marcus - ICCBL user agreements)

Removed, as the trunk is now used for collaboration

Added 'jar' target to build.xml, allowing Screensaver classes to be shared with other projects that depend upon this project.

Belongs to previous commit

In the cellHTS2 report the following elements were not generated:
1. heatmaps for intensities per plate and replicate
2. Screen-wide image plot of the scored values
Ad 1: Due to the setting of the write report argument plotPlateArgs = FALSE
Change it to TRUE
Ad 2: Due to lacking of the "scored" element in the cellHTS2 parameter of the writeReport.
Have added it.
I also included changes to the R and JUnit test voor writeReport. 

added a new user checklist item

set default plate size to 384 well plates

bug fixes and unit test updates

Continued work on PlateSize refactoring.  Checkpointing now that initial implementation is complete.

In R 2.7.0 png needs cairo >= 1.2, however on production system is 1.0.2 installed. 
installing new library will probably create dependency issues
For now problem solved by overloading png function with default type="cairo1"

Moving release-1.6.1-maint branch under /branches/iccbl, to improve repository organization

Moving release-1.5.3-maint branch under /branches/iccbl, to improve repository organization

Removing aborted-popup-help branch, as work on this branch is no longer needed.

Removing schema-upgrade-2007 branch, as work on this branch has been completed and reintegrated as of 2007-10-05, rev 1928.

Removing search-result-redesign branch, as work on this branch has been completed (2008-02-19, rev 2188).

Comprehensive support for multiple plate sizes, by introducing PlateSize enum to encapsulate details of plate extents (rows and columns).  All constants related to plate size are replaced with PlateSize type.  This is follow-up work for [#839] ("Support for 96 wells plates").  As of this revision, first draft of code changes are complete, but test code needs to be updated, and migration script added for Library.plateSize property.

* merge from collaboration branch from 2969 to 2970
* add some other documentation

Made running the cellHTS2 writerReport function configurable for the parameter progressReport. WriteReport runs default with a progress window which requires the user to have access to the display.  Running it with progressReport= FALSE, no progress window is shown, which on a server is obsolete anyway. Now possible to run Rserve under account without access rights to display. Created new property in the screensaver.properties file: cellHTS2report.writeReport.progressReport=FALSE. Alternatively set to TRUE.

See for example cfg/bii/screensaver.properties.bii.sample 

merge http://forge.abcd.harvard.edu/svn/screensaver/branches/trunk revision to http://forge.abcd.harvard.edu/svn/screensaver/branches/collaboration head revision

Added documentation

In ScreenResultviewer the check if file exists is now correctly working in combination with the feature of full path for cellhts output directory.

CellHTS2 output can now be written to location outside the webapps directory. The root directory for the output is now in stead of webapps, defined in the property cellHTS2report.filepath.base in the screensaver.properties. See example cfg/bii/screensaver.properties.bii.sample. The data retrieval is done by a new class DataServlet in the cellHTS2 package. The writing and reading is working, however the presenting of the url on the screen isnot yet. 

Add CreateImportXls

fixed imports

Resolved [#1692] ("reset" search command in wells browser returns all wells).  Reset buttons on individual table columns now cause WellSearchResults to revert to empty result if reset causes all columns to have undefined criteria.

 [#1656] well search results reset domain restriction on search action

 [#1627] well search across all libraries
 add domain restriction so that "retired" libraries do not show up in the searches

fixed ticket links in changelog.html and newsViewer.xhtml

updates for 1.7.3

[#1196] reagent viewer annotations information
formatting for said information

[#1683] "Download SD File" should not be exposed in reagent viewer

updated screensaver-example.sql.gz database snapshot for 1.7.3 release

updated version for 1.7.3 ICCB-L release

[#1681] error on opening reagent viewer
remove references to isSpanCell

updated version for 1.7.3 ICCB-L release

Resolved [#1680] ("browse wells" "reset all" command should reset to empty search result)

[#1675] user viewer checklist item "N/A" command error
apply logic from the maintenance branch fixes (key off the checklist item entry not the row index)

Resolved [#1679] ("browse wells" switch search fields location causes full well search result).  

[#1196] reagent viewer annotations information
- exclude the summary information to aid in UI readability

volume & concentration fields now converted to "reasonable" units after editing activity or cpr (avoids surprising the user, as this conversion will be done anyway the next time the parent entity is loaded/viewed).

- allow values of 0 in UI for volume and concentration type
- part of this was implemented by just removing the validator in activity viewer for concentration field; this is okay for now, since we don't yet have validators for other volume/concentration fields, and we need to devise a consistent implementation

- minor fixes for handling of null values for volume and concentration fields in ActivityViewer and CPRViewer, to allow null values to remain null after save
- minor fix to ActivityViewer to reset volume and concentration fields on viewing new activity

- fixed error in 2822.sql
- as the effect of this fix was already apparent in ICCB-L databases, the error was presumably found and fixed during an earlier release process, but never committed

Library Viewer:
  - disabled Delete command for now, due to exceedingly bad performance
  - shows description and vendor fields, even if empty (to be consistent with the rest of the UI)
  - prevent editing of start/end plate after initial library creation, since code is not yet in place to update the library set of wells in response in to changes in these fields

[#1594] display library 'screening status' field
 - also in the library browser

Made Library.screeningStatus editable in Library Viewer, as follow on work to [#1594] ("display library 'screening status' field")

[#827] merge library create, edit, delete functions from IMCB
adjust the styleClass for the  'screen type' and 'library type' fields

fixed a unit test, by adding 'messages' bean to the spring test context (not ideal, but pragmatic)

removed unused message

fixed LazyInitEx when ReagentSearchResults used in searchAllReagents() mode (currently, only used by unit tests)

- added unit test for recently updated WellSearchResults.searchAllWells()
- fixed existing unit tests that made use of this method

- merged 1.7.2 maintenance release changes into branches/iccbl/trunk, for 1.7.3 release (merge -r2879:HEAD http://forge.abcd.harvard.edu/svn/screensaver/tags/iccbl-releases/release-1.7.2-maint-1)
- merges [#1592] and [#1593]

- Resolved [#1662] (activity viewer "performed by" incorrect if user no longer a valid selection).  Defect also affected the CPR Viewer's "requested by" field.
- Code that determines "candidate" selections for these associations now also allows the *existing* selection to be considered valid, even if it would not be considered valid if the parent entity was created anew.  This is the correct solution, as it accommodates the historically valid candidates.
- Moved the code that determines "candidate" selections for Activity and CPRs into the domain model classes, since this is domain model concern, not a UI concern.
- Fixed bug in CPRV, introduced during [#1617], causing error when viewing some CPRs (lazy init'ed data members were accessed directly)

fixes to 2925.sql schema update script

[#1661] search results for "quick" searches should sort ascending on entity ID

 [#1656] well search results reset domain restriction on search action

[#1660] add 'data sharing agreement signed date' field to user
-actually implemented as a new checklist item

 [#1562] Add lab affiliation category column to user search results
 - added a LabAffiliationCategory Column that may be seaarched

Fixed long-standing cosmetic defect in menu panel, re-introducing whitespace before the build number that was lost during Facelets conversion.

Unit test for [#1601] "cpr should remove duplicate lcps", testing that LibraryPoolToDuplexWellMapper throws BusinessModelViolationException if the pool wells of multiple SCPs happen to map to the same duplex wells.

Minor updates to generic error message in messages.properties.  Removed unused messages.  Changed text of applicationError and businessError.  Added usage comments.

Fix bug in CPRViewer where volume fields were not being reset before viewing a new CPR.

Resolved [#1601] "cpr should remove duplicate lcps"
  - LibraryPoolToDuplexWellMapper now throws BusinessModelViolationException if the pool wells of multiple SCPs happen to map to the same duplex wells.
  - Also, for defensive programming, CherryPickRequestAllocator now throws DataModelViolationException if LCP source wells are not unique.

[#827] merge library create, edit, delete functions from IMCB

[#827] merge library create, edit, delete functions from IMCB
[#1594] display library 'screening status' field

[#827] merge library create, edit, delete functions from IMCB

sql files were not transferred with earlier merge from collaboration branch.

Resolved [#1645] ('browse users' initial sort order)

Resolved [#1617] "cpr viewer lcp and scp data tables are not user independent".  Reworked design so that _{lab,screener}SearchResults are reinitialized when a
new CPR is to be viewed, forcing dynamic rebuild of table columns, and so are no longer defined as static data members.

deleted 2900 and 2901 schema update scripts, since they have a no-op net effect.

removed unused imports

removed unused imports

[#1645] 'browse users' initial sort order

[#1651] well viewer look & feel broken

added new class file that was missed when creating iccbl/data-table-refactor branch

Refactor the data table and search result framework

[#1605] new activity viewer replicate fields

[#1613] Create a link from each Annotation in the Reagent View to the Study

 [#1607] billing info field update
 - revert the changes to BillingInformation related to earlier commits for this ticket
 - apply changes to BillingItem as appropriate for this ticket

 [#1607] billing info field update
 - revert the changes to BillingInformation related to earlier commits for this ticket
 - apply changes to BillingItem as appropriate for this ticket

 [#1607] billing info field update
 - revert the changes to BillingInformation related to earlier commits for this ticket
 - apply changes to BillingItem as appropriate for this ticket

[#1607] billing info field update - make sure the field is nullable and name is consistent with UI usage.

[#1427] add concentration field to cpr - enable for RNAiCherryPickScreening

[#1627] Well Search across all libraries and studies

[#1627] Well Search across all libraries and studies

formatting

[#1584] reset search fields on menu
Upon successful search, these fields are reset, following the pattern for the screener search field.

[#1604] activity viewer field rename
"Volume Transferred" --> "Volume Transferred per Replicate"

[#1607] billing info field update

[#1515] screen result 'last modified date' should be updated each time a new result is uploaded.  
Updated the test and javadocs.

[#1515] screen result 'last modified date' should be updated each time a new result is uploaded.

[#1294] How to display retired library copies in search views

[#1558] volume fields: separate value and units

fix error in displayed volume value (actually showing concentration value)

[#1196] reagent viewer annotations information
show meta-info for the annotations

for library create

null checks for values

set the summary also

Partial, and possibly temporary, fix for [#1593] (add funding support ‘Marcus Library Screen’)

Resolved [#1592] (user checklist item updates not respecting user-specified dates).  Broke when checklist item groupings were added, since row indexes were used to determine the user-specified date, but row indexes were group-local.  now using checklistItem as key for a user-specified date map.

merged changes inadvertently made to tags/iccbl-releases/release-1.7.2 to branches/iccb/release-1.7.2-maint (merge -r2860:HEAD http://forge.abcd.harvard.edu/svn/screensaver/tags/iccbl-releases/release-1.7.2).  worth doing, since we have more maint fixes forthcoming.

ICCB-L production release 1.7.2 maintenance branch

fix bug where "requestedHarvardIdExpirationDate" was being set to "harvardIdExpirationDate", see [#1373] add 'requested havard id expiration date' property to user and user viewer

allow entry & proper display of null values, for:
[#1427] add concentration field to cpr
[#1426] add concentration field to library screening activity
[#1558] volume fields: separate value and units (partial fix)

[#1427] add concentration field to cpr
[#1426] add concentration field to library screening activity
[#1558] volume fields: separate value and units (partial fix)

log message

Reintroduced explicit Concentration type, which is now a subclass of Quantity.  Improves code readability, replacing 'Quantity<Concentration>' with just 'Concentration'.  This is follow-up work on [#1512] (domain model support for concentration units).

fix help to show correct identifiers for "screen type" and "study type"

added comment explaining fix for [#1361]

Minor fixes for [#1276] checklist item "not applicable".  
  - UI now only shows "N/A" command if checklist event history is empty
  - schema migration script works w/existing data in check_list_item_event
  - updated data model constraints for adding n/a checklist item event

for [#1361] edit well volumes: new volume values disappear on page scroll; change immediate to false on datascroller

fixed comments

[#1276] checklist item "not applicable" - For a given user, a checklist item should be able to be flagged as "not applicable", permanently disabling further activation/deactivation/completed events from being recorded for that checklist item.

add checklist_item_group column

changes for [#827] create library from UI - merge "create library" functionality from imcb branch

changes for [#827] create library from UI - merge "create library" functionality from imcb branch

changes for [#827] create library from UI - merge "create library" functionality from imcb branch

changes for [#827] create library from UI - merge create library functionality from imcb branch

changes for [#827] create library from UI - merge create library functionality from imcb branch

Fixed bug with CPR Viewer warnings not being issued when cherry picks are initially added (CPR was not flushed to database, causing HQL queries to not consider the new cherry picks)

- Resolved [#1570] (error on adding cherry picks to cpr).  
- HQL query now used to count distinct compounds for calculating Compound CPR "cherry pick allowance used" value, rather than loading all screen result wells and compounds into the Hibernate session (which did not scale).
- Refactored "cherry pick allowance policy" implementation into an interface and explicit "policy" classes (edu.harvard.med.iccbl.screensaver.policy package), allowing these policy implementations to be spring-configurable.

Default screening status for a new library is now set to 'Allowed', fixing NPE (screeningStatus prop is non-nullable).

refactor Concentration type to use Quantity type

Added MolecularInfoDumper utility, for investigation of [#1554] (molecular weight and formula ignoring implicit H atoms).

update changelog.html, newsViewer.xhtml for 1.7.2 final tickets

ConcentrationTest

not required

add showMessageForLocalCompenentId

allow null concentrationvalue

[#1426] add concentration field to screen
[#1512] domain model support for concentration units
[#1427] add concentration field to cpr

customize information for IMCB

reorder data table's visible columns [#1471]

adjust value of CONNECT_TIMEOUT

Integrate Annotator in Screensaver

Implement email notification for library content import.
Threading is used here.

servlet filter for cellHTS2 report authorization [#882]

Implement 'display control well information' feature.
Control well information is:
- uploaded with screen result file
- displayed in screen result data table

 - use Andrew's java package for cellHTS2
 - cellHTS2 report configurable from screensaver.properties file
 - servlet filter for cellHTS2 report authorization [#882]
 - help for cellHTS2 [#1472] [#1475]

'hide compound' feature [#1470]

update help [#1475] [#1472]

help for control well information upload [#1475]

help for control well information upload [#1475]

The following are 
1. display control well information columns
2. display assay well type [#1473]
3. implement 'hide compound' feature [#1470]

update comment

Temporarily removing the "Filtered View" title on UserSearchResults, until we determine a way to show this title conditionally (based upon whether search criteria exist).  Previously, the "Filtered" title remained, even if search criteria were reset.

Fixed screen result viewer 'Data' and 'Data Headers' collapsible panels.