Adding Format Identification Tools

From Archivematica
Jump to navigation Jump to search
Warning: These instructions are out of date as of Archivematica 1.0. Please see the official Archivematica documentation.

This will show a developer how to add a new tool new tool to identify file formats (file ids), and allow normalization based on the newly identified file formats.

The workflow this section is looking to implement are:

  • Make the fileID tool a selectable choice
  • Specify that normalization will be using that fileIDType
  • Identify files using that tool, with valid archivematica fileIDs (format IDs)
  • Use those format IDs and associated commands to normalize files.

Since each link in the workflow depends on the next one, we need to start at the end and work our way back.

All pk's (primary keys) in the database are UUIDs generated with 'uuid -v4' on the command line

Add tool[edit]

To include a new tool in the archivematica packages, it should be a dependency package itself.

FileIDTypes[edit]

First, an entry for FileIDTypes is required for the new tool. 

INSERT INTO 
    `FileIDTypes` (`pk`, `description`, `replaces`)
VALUES 
    ('afdbee13-eec5-4182-8c6c-f5638ee290f3','FileIDByFIDO',NULL);

Identify Files Script[edit]

The script to identify the file must be created, and an entry for it added to the FilesIdentifiedIDs table. For Fido this script is archivematicaFido.py All the arguments the script accepts will be listed as arguments in StandardTasksConfigs, below.

This script must also be added to the client as a supported module by adding it to archivematicaClientModules. Please keep the list in alphabetical order. 

archivematicaFido_v0.0 = %clientScriptsDirectory%archivematicaFido.py

Run tool[edit]

Now, we set up the chain link that will run the script.

FIDO workflow link example: 

-- Set up the UUIDs to be used as pks
SET @TasksConfigPKReference = '46883944-8561-44d0-ac50-e1c3fd9aeb59';
SET @TasksConfig = '7f786b5c-c003-4ef1-97c2-c2269a04e89a';
SET @MicroServiceChainLink = '4c4281a1-43cd-4c6e-b1dc-573bd1a23c43';
SET @MicroServiceChainLinksExitCodes = 'd7653bbd-cd71-473d-b09e-fdd5b36a1d65';

--Where to go once completed is the same for both
SET @YLink = '83484326-7be7-4f9f-b252-94553cd42370';
SET @defaultNextChainLink = @YLink;
SET @NextMicroServiceChainLink = @YLink;

INSERT INTO 
    StandardTasksConfigs (pk, filterFileEnd, filterFileStart, filterSubDir, requiresOutputLock, standardOutputFile, standardErrorFile, execute, arguments)
VALUES 
    (@TasksConfigPKReference, NULL, NULL, 'objects/', FALSE, NULL, NULL, 'archivematicaFido_v0.0', '--fileUUID "%fileUUID%" --SIPUUID "%SIPUUID%" --filePath "%relativeLocation%" --eventIdentifierUUID "%taskUUID%" --date "%date%" --fileGrpUse "%fileGrpUse%"');
  • This creates the configuration for running the script
  • filterFileEnd, filterFileStart, filterSubDir: What files to run the identification script on. In this case, all the files in the objects/ subdirectory
  • execute: The script to execute, as it was defined in archivematicaClientModules
  • arguments: Anything that should be passed on the command line to the script. It should include both the argument, and the value to pass to the argument.
INSERT INTO 
    TasksConfigs (pk, taskType, taskTypePKReference, description)
VALUES
    (@TasksConfig, 'a6b1c323-7d36-428e-846a-e7e819423577', @TasksConfigPKReference, 'Identify file formats with FIDO');
  • This maps between the task (inserted below), and its configuration
  • taskType: We want this to run on every file, and a6b1c323-7d36-428e-846a-e7e819423577 is the pk for that
  • taskTypePKReference: is semantically a foreign key to StandardTasksConfigs, where our config is stored
INSERT INTO 
    MicroServiceChainLinks (pk, microserviceGroup, currentTask, defaultNextChainLink)
VALUES 
    (@MicroServiceChainLink, @microserviceGroup, @TasksConfig, @defaultNextChainLink);
  • This creates the actual link to run
  • microserviceGroup:
  • currentTask: Points to the TasksConfigs we just inserted
  • defaultNextChainLink: where to go if we can't figure out where else to go, usually on failure
INSERT INTO 
    MicroServiceChainLinksExitCodes (pk, microServiceChainLink, exitCode, nextMicroServiceChainLink)
VALUES 
    (@MicroServiceChainLinksExitCodes, @MicroServiceChainLink, 0, @NextMicroServiceChainLink);
  • This configures where to go once the MicroServiceChainLink is complete, based on exit code
  • microServiceChainLink: the chain link we just defined
  • exitCode: the exit code this is valid for. In this case, 0=success. These should obviously match what the script generates
  • nextMicroServiceChainLink: where to go next. The last link, to return to the processing workflow already exists, and must be defined as the next chain to process. The link is 83484326-7be7-4f9f-b252-94553cd42370
SELECT 
    MicroServiceChainLinks.pk, TasksConfigs.description 
FROM 
    MicroServiceChainLinks 
    JOIN TasksConfigs ON TasksConfigs.pk = MicroServiceChainLinks.currentTask 
WHERE 
    MicroServiceChainLinks.pk='83484326-7be7-4f9f-b252-94553cd42370';
+--------------------------------------+---------------------------------------------------------------+
| pk                                   | description                                                   |
+--------------------------------------+---------------------------------------------------------------+
| 83484326-7be7-4f9f-b252-94553cd42370 | Resume after normalization file identification tool selected. |
+--------------------------------------+---------------------------------------------------------------+

Add workflow[edit]

Set Selection[edit]

The first step in the workflow is to set the selection as the tool to use during normalization. This is done by making an insert into the unit's variables table for the variable normalizationFileIdentificationToolIdentifierTypes. The value set is a piece of a SQL query used in linkTaskManagerSplitOnFileIdAndruleset.py to restrict the fileIDs used to the desired type.

For FIDO: 

INSERT INTO 
    `MicroServiceChainLinks` (`pk`, `currentTask`, `defaultNextChainLink`, `defaultPlaySound`, `microserviceGroup`, `reloadFileList`, `defaultExitMessage`) 
VALUES 
    ('982229bd-73b8-432e-a1d9-2d9d15d7287d','1e516ea6-6814-4292-9ea9-552ebfaa0d23','4c4281a1-43cd-4c6e-b1dc-573bd1a23c43',NULL,'Normalize',1,'Failed');

INSERT INTO 
    `TasksConfigs` (`pk`, `taskType`, `taskTypePKReference`, `description`) 
VALUES 
    ('1e516ea6-6814-4292-9ea9-552ebfaa0d23','6f0b612c-867f-4dfd-8e43-5b35b7f882d7','f130c16d-d419-4063-8c8b-2e4c3ad138bb','Set SIP to normalize with FIDO file identification.');

INSERT INTO 
    `TasksConfigsSetUnitVariable` (`pk`, `variable`, `variableValue`, `microServiceChainLink`) 
VALUES 
    ('f130c16d-d419-4063-8c8b-2e4c3ad138bb', 'normalizationFileIdentificationToolIdentifierTypes', 'FileIDTypes.pk = \'afdbee13-eec5-4182-8c6c-f5638ee290f3\'', NULL);

INSERT INTO 
    `MicroServiceChainLinksExitCodes` (`pk`, `microServiceChainLink`, `exitCode`, `nextMicroServiceChainLink`, `playSound`, `exitMessage`) 
VALUES 
    ('82c97f8d-087d-4636-9dd9-59bbc04e6520','982229bd-73b8-432e-a1d9-2d9d15d7287d',0,'4c4281a1-43cd-4c6e-b1dc-573bd1a23c43',NULL,'Completed successfully');

And the chain that starts running the chain links we defined above: 

INSERT INTO 
    MicroServiceChains (pk, startingLink, description)
VALUES 
    ('c76624a8-6f85-43cf-8ea7-0663502c712f', '982229bd-73b8-432e-a1d9-2d9d15d7287d', 'FIDO');
  • MicroServiceChains.startingLink: the chain link we defined above
  • MicroServiceChains.description: name of the tool

Chain Choice[edit]

The choices for file identification tool choices are made at link f4dea20e-f3fe-4a37-b20f-0e70a7bc960e.

Additional choices can be added by adding entries. 

SELECT 
    chainAvailable, startingLink, description 
FROM 
    MicroServiceChainChoice 
    JOIN MicroServiceChains ON chainAvailable = MicroServiceChains.pk 
WHERE 
    choiceAvailableAtLink = 'f4dea20e-f3fe-4a37-b20f-0e70a7bc960e';
+--------------------------------------+--------------------------------------+---------------------+
| chainAvailable                       | startingLink                         | description         |
+--------------------------------------+--------------------------------------+---------------------+
| 229e34d9-3768-4b78-97b7-6cd4a2f07868 | b549130c-943b-4791-b1f6-93b837990138 | extension (default) |
| c44e0251-1c69-482d-a679-669b70d09fb1 | 56b42318-3eb3-466c-8a0d-7ac272136a96 | FITS - DROID        |
| 1d8836cf-ac02-437c-9283-4ddb7b018810 | 37f2e794-6485-4524-a384-37b3209916ed | FITS - ffident      |
| d607f083-7c86-49a2-bc36-06a03db28a80 | 766b23ad-65ed-46a3-aa2e-b9bdaf3386d0 | FITS - JHOVE        |
| 586006d1-f3af-4b5f-9f1a-c893244fa7a9 | d7a0e33d-aa3c-435f-a6ef-8e39f2e7e3a0 | FITS - summary      |
| 50f47870-3932-4a88-879d-d021a24758ad | f87f13d2-8aae-45c9-bc8a-e5c32a37654e | FITS - file utility |
| c76624a8-6f85-43cf-8ea7-0663502c712f | 982229bd-73b8-432e-a1d9-2d9d15d7287d | FIDO                |
+--------------------------------------+--------------------------------------+---------------------+

Finally, we add the actual choice: 

INSERT INTO 
    MicroServiceChainChoice (pk, choiceAvailableAtLink, chainAvailable)
VALUES 
    ('e95b8f27-ea52-4247-bdf0-615273bc5fca', 'f4dea20e-f3fe-4a37-b20f-0e70a7bc960e', 'c76624a8-6f85-43cf-8ea7-0663502c712f');
  • MicroServiceChainChoice.choiceAvailableAtLink: MicroServiceChainLink where the choice is made available. We know this is f4dea20e-f3fe-4a37-b20f-0e70a7bc960e
  • MicroServiceChainChoice.chainAvailable: The chain we just added

Add FPR rules[edit]

Populating the local FPR rules can be done by creating a temporary script, and using a set of sample files with extensions. Use the FileIDs for the extensions to populate values for already in preservation/access status, and related commands.

Fido example: [1]

Retrieved from "https://wiki.archivematica.org/index.php?title=Adding_Format_Identification_Tools&oldid=13314"