funderburkjim
commented
4 years ago pdffiles.txt
When a display wants to present the scanned image of a particular page in a particular dictionary (such as page 346 for MW), it consults the 'pdffiles.txt' file for that dictionary. All these files may be found in this csl-websanlexicon. For the mw example, the file is csl-websanlexicon/v02/distinctfiles/mw/web/webtc/pdffiles.txt.
This file has 3 colon-delimited fields.
- identifier -- more on this in the later
- the filename of the image. The full URL of this image varies, depending on the display
program. But this filename is the last part of the URL.
- The filename suffix is usually 'pdf', but for some dictionaries it is 'png' or 'jpg'. See table in next comment.
- a 'title' for the file.
- for the normal body page, this is usually the first word on the page
- for 4 dictionaries (pw, stc, wil, gra), this field is absent.
- for some dictionaries, it may be present but as an empty string.
- this 'title' field is unused by display programs.
servepdf.php
The getfiles function within servepdfClass.php in csl-apidev repository shows how the identifier field within the lines of pdffiles.txt is used to get the filename for the body pages.
This is in response to an inquiry from @vocabulista . He has the need to rename all the scanned image file names in a consistent, flat, way.
First, the scanned images for each dictionary can be separated into two piles:
The following comments pertain only to the first pile.