Preparing Collections on the S Drive for Online Delivery and Storage
- 1 Things to do before marking a folder as "Store"
Things to do before marking a folder as "Store"
The collection number u0003_0000001 will be used as an example.
Check Subfolders of Collection Level Folder
The Collection Level folder contains subfolders and their content must adhere to certain specifications prior to the collection being considered ready to "ship" for online access and long term storage.
- This folder must exist.
- Must contain:
Make sure to refer to the Collection_Information page regarding acceptable data values.
Include a text version of the log file with every batch; previous versions of the log will be overwritten by the newest version. Specifically, since this log file's source file is an Excel workbook, it is the "log" sheet within the Excel file that needs to be saved as u0003_0000001.log.txt. This is the sheet with all the scanning data (technician, dates, # of scans, etc.).
- May also contain:
For example, audio collections often have significant item-level notes that we want to retain. These plain text files can be saved with a ".notes.txt" extension - i.e. "u0008_0000001_0000001.notes.txt".
- If multiple Digital Collections spawn from the same Analog Collection, there can be more than one Collection Information XML file as follows:
- This folder must exist.
This is tab-delimited text export of the original spreadsheet. (If this file was exported out of Excel as a tab-delimited txt file, you must open the file with Notepad ++ and do a Search and Replace to remove the quotation marks that have been inserted at each tab.) The source .xlsx spreadsheet should be moved to S:\Digital Projects\Administrative\collectionInfo\forMDlib\needsRemediation. If this is a large or ongoing collection, the tab-delimited text export should contain ONLY the metadata for the items currently being transported to storage. The text file itself should have a period and then a number to indicate which portion, or "batch", of the complete metadata this is. The first tab-delimited export would be named, for example, u0003_0000001.1.txt, and would contain the first 500 entries, for example. The second tab-delimited export, for items 501-1000, would be named u0003_0000001.2.txt, and so forth. Thus, only by collecting all these tab-delimited exports do we have a complete set of descriptive metadata for the collection items.
For more on how to parse these "batches" out from the complete set of descriptive metadata, see Parsing Metadata.
- a MODS folder
- This folder will contain all the MODS files created via Archivist Utility (see: Making MODS).
- May also contain:
All should use ANSI or UTF-8 (without BOM) encoding.
If you are comfortable with XML, please open the EAD file and look for this line (should be the ~4th one down): <eadid countrycode="US" mainagencycode="US-US-ALM"></eadid> If the collection number is not there (what we name the file: u0003_0000580) then please enter it, so that line looks like this: <eadid countrycode="US" mainagencycode="US-US-ALM">u0003_0000580</eadid> This way the file self-references and can be found by this number during searches if we index it properly. Also, if something gets misnamed somewhere, this will help to sort out the problem.
- This folder must exist.
Note: we may break Scans folders into chunks for manageability, for more information click here.
- Must only contain:
- scans (tiffs/wavs) of non-compound objects and compound objects (inside respective subfolders)
All other files types will not be retained. Temporary files and thumbs.db files do not have to be deleted since they will be removed upon transfer to Storage.
- This folder must exist only if transcripts exist.
- Must only contain one or more of the following types of files:
- u0003_0000001_0000001.tif, u0003_0000002_0001.tif, etc. - corresponding to non-compound and compound objects (inside respective subfolders).
- u0003_0000001_0000001.txt, u0003_0000002_0001.txt, etc. - plain .txt files corresponding to non-compound and compound objects (inside respective subfolders).
- u0003_0000001_0000001.ocr.txt, u0003_0000002_0001.ocr.txt, etc. - plain .txt files of OCRed tiffs corresponding to non-compound and compound objects (inside respective subfolders).
If cleaned up .txt files exist, remove the corresponding .ocr.txt file.
Perform Quality Control Tasks
Note: quality control tasks will already have been performed by scanning technicians beforehand, but it's a good idea to run the QC scripts again. This takes very little time.
Spot check all .xml, .txt, and .xlsx files
- Check all such files for proper filenames and extensions.
- Open all such files and look for anomalies and inconsistencies, misspellings, and missing data, etc.
- make sure fields such as Funder(s), Funding Information, Repository Collection, Digital Collection, Digital Publisher, etc. are filled out.
The required fields should have a green header cell in the descriptive metadata to aid in verifying the presence of required content.
- ideally, no additional fields such as "Notes" are in the Metadata file. "Notes" as such should be deleted or moved to the appropriate row in the log.txt file.
- make sure the Format column in the Metadata file has not been altered to the Time format as it sometimes is:
If a tab delimited metadata file is opened via Excel (especially by right clicking the file and choosing to open in Excel), the format column if like: 3 p., 4 p., etc. Will get interpreted as 3:00 PM, 4:00 PM, etc. If then resaved as .txt, times will have been saved instead of page #s. The way around this is to have Excel open first, choose Open. open your text file and while you are being interrogated by Excel about how to import, set the Format column as "Text".
Obviously, if errors are found *after* text exports and MODS files are made, then the Excel file needs to be corrected and the text and MODS files remade.
Check all Folder names
- Make sure folders are named correctly and that there are no superfluous word concatenations to object level folders, etc.
Match Data across documents and folders
This table attempts to show how data in one of our documents/folders should match with data in another document/folder.
Corresponding data must properly match/equate prior to marking "Scans" folders as "Store".
|TrackingFileNames||Admin XML||Metadata||Archivist Queue (\"Selection.xlsx\")||TrackingFiles||Finding Aid||Scans folder|
|Collection Number||[first 14 characters of Filename]||Collection Number|
|physical location||Manuscript_Number [Is this true?]||Manuscript Number; Physical Location||MSS #|
|primary analog format||Genre [since there can be multiple genres in the metadata, this will correspond, in average, to the TrackingFileNames Primary Analog Format]||Genre/type|
|source collection||Analog_Collection_Name||Repository Collection||Name of Analog Collection||Title|
|Description||Digital_Collection_Description||Blurb||Abstract; or Scope and Contents [limited to digitized portion]|
|Digital_Collection_Name||Digital Collection||Project Name|
|Type_Of_Content||Type(s) [since there can be multiple Types in the Metadata, this will correspond, in average, to the Admin XML Type_of_Content]||Genre/type|
|Finding_Aid_Link||Link to finding aid|
|Total Scans (Scans + Transcript Scans)||[use PERL script or folder search to retrieve number of Total Scans (tifs or wavs)]|
|[count total objects in Metadata sheet]||Total Objects||[use PERL script or folder search to retrieve number of objects (singletons + compound object subfolders)]|
From a digital preservation/delivery perspective, it's not as important to match information to the Archivist Queue spreadsheet, though it would be ideal if possible. Also, it isn't always feasible to match the number of actual scans vs. what is notated in the TrackingFiles, although that is also ideal.
Moving the Files to the Server
The folder should now be prepared to run scripts and place the files on the storage server.