This script is post-processing minutes that have been produced by scribejs. The goal is usually to extract information from the minutes that may make the life of Working Groups easier…
At this moment, there are two such post-processing steps: extract the resolutions into a separate file, and add a notice to each issue that was discussed during a call. See the details below.
The script has been developed in TypeScript; the “compiled” files are in the ./dist directory. The script can be run by:
Usage: node dist/main.js [configuration_file]
The configuration files are shared with the ones used by scribejs to avoid repeating the data, although many of the data in the configuration file is therefore ignored by the post-processing steps. See the description configuration for more details.
The script avoids unnecessary GitHub access, and duplication of data or issues, by looking at a specific file (whose location is in the configuration file) that lists the minutes already “processed”. That JSON file is of the form:
{
"date" : // Date of the last generation,
"file_names" : [
// Array of the file names for the minutes that have already been processed
],
// other types of data
}
and is updated at each run after adding the new minutes that have been processed, if applicable. An example for the generated JSON file is on the EPUB WG’s site.
scribejs adds the references to resolutions accepted during the call into the JSON-LD metadata. This steps extracts these data and adds them to the generated processing file (see the example above). The resulting file structure is as follows:
{
"date" : // Date of the last generation,
"file_names" : [
// Array of the file names for the minutes that have already been processed
],
"resolutions" : [
{
"year" : // year of the meeting,
"date" : // date of the meeting,
"number" : // index of the resolution for that meeting
"text" : // text of the resolution, formatted in HTML
"url" : // pointer to the minutes where the resolution was discussed
"call" : // the name, i.e., the IRC channel name, for that meeting
}
... // etc. (and entry for each resolution)
]
}
The resulting file can be used by a client-side script to display the list of resolutions. See, for example, the EPUB WG deployment, which uses the client-side script generated from the separate Typescript module in this repository.
scribejs adds the references to actions raised on the call into the JSON-LD metadata. This step extracts those actions and raises special issues on GitHub (the information stored in the metadata includes the reference to the GitHub repository).
The configuration file contains a separate flag to switch this behavior on and off for each repository.
scribejs adds special “directives” into the minutes (relying on a scribe feature) when a section (or subsection) discusses specific issues or pull requests. These directives are of the form:
<!-- scribe owner/repo/issuenumbers owner/repo/issuenumber2 -->
The post-processing steps extracts the (sub)sections that contain these directives, creates a notice referring to the minutes, and the possible resolutions, and finally adds a comment to the issue on GitHub. This, effectively, “binds” the WG discussions to the issues and vice versa. The configuration file contains a separate flag to switch this behavior on and off for each repository (some WG-s do not systematically separate the issues into subsections, which would make this processing step meaningless…).
This post-processing step should be invoked using the -m flag, with a possible -d date (in ISO format; defaults to today's date). To avoid spamming mailing lists, this option means that only the email action is executed.
Maintainer: @iherman