Collect availability responses (mini-Doodle/WhenIsGood), backed by Grist

Users arrive via a personalized link (/availability/<key>), created in a table of "Availability requests" in Grist. Users are shown a calendar, allowing them to

• draw their availability over a shown set of ranges,
• select yes/no for predefined time slots, and
• submit a textual response.

Data from their response is once again recorded in Grist.

Needed configuration: see the dev script for needed environment variables.

To run, first run poetry install (see poetry docs), then run dev.sh. Generalize/deploy from there.

Schema assumptions

Input (mostly)

Availability_requests needs columns:

• Request_group: (int/any) Used to find request timespans. In my case, a reference to another table, Availability_request_groups, but it need not be.
• Key: (text) Used as part of the link to the personalized calendar.
• Person: (any) Identifier of person to whom the request is addressed. Copied verbatim into Availability.
• Name: (text) Name of the person or entity whose availability is being requested.
• Allow_maybe: (bool, optional) Whether timespans may be marked as 'If I must'.
• Message: (text) A message shown to the user on their personalized calendar.
• Min_span_minutes: (number, optional) The minimum number of minutes required in each time span supplied by a respondent.
• Responded: (unix timestamp, input/output) To record whether (and when) the user has responded to this request.
• Response: (text, output) To record the user's textual response.

Request_timespans needs columns:

• Request_group: (int/any) Identifier of the group of requests, matches Availability_requests.Request_group.
• Start: (unix timestamp)
• End: (unix timestamp)
• Allow partial: (bool) If yes, will draw a timespan in the background, allowing the user to draw their availability over it. If no, only allows the user to say yes/no to each timespan.

Contextual_availability_requests (optional) has columns:

• Request: Reference to Availability_requests. The request for which context is being provided.
• Description: A description of the contextual relationship (e.g. "student"/"instructor").
• Contextual_request: Reference to Availability_requests. The request from which contextual data should be shown.

Output

Availability:

• Request_group: (int/any) Identifier of the group of requests, matches Availability_requests.Request_group.
• Person: (any)
• Start: (unix timestamp)
• End: (unix timestamp)
• Available: (bool)
• Maybe: (bool) (only used if Allow_maybe is present and truthy)
• Request_timespan: (reference to Request_timespan) Added if Allow partial is false for a given span.

UWSGI config for deployment

[uwsgi]
plugins = python311
socket = /tmp/uwsgi-grist-av.sock

env = GRIST_ROOT_URL=https://grist.tiker.net
env = GRIST_API_KEY_FILE=/home/grist-av/.grist-api-key
env = GRIST_DOC_ID=rLJPGJ9RLJ4TRVx4AxT2tW
env = SECRET_KEY=CHANGE_ME
env = CAL_TIMEZONES=America/Chicago,local,UTC

# Optional. Only effective if both are provided.
env = NOTIFY_FROM=andreask@illinois.edu
env = NOTIFY_TO=andreask@illinois.edu

chdir = /home/grist-av/grist-availability
module=availability.app:app
uid = grist-av
gid = grist-av
need-app = 1
workers = 1
virtualenv=/home/grist-av/grist-availability/.venv
buffer-size = 16384