[Jungfrau][Correct] Account for missing fixed gain constants when burst mode in dynamic gain

Description

MID started recorrecting data from October in burst mode. These data were corrected before with dynamic gain constants and no fixed dark constants were taken.

This MR accounts for this by expanding https://git.xfel.eu/calibration/planning/-/issues/196 to force fixed gain constants for burst mode in dynamic gain. Then in case of no available constants check for dynamic gain constants. These constants will be accepted only if they don't have more than 3-day time difference with the RAW data.

• New functions: jungfrau_cal_mdata and force_fixed_gain_constants
• Try retrieving fixed gain constants, try retrieving dynamic gain constants if no fixed constants
• Dont accept dynamic gain constants older than 3 days

How Has This Been Tested?

• TBD

Relevant Documents (optional)

This is an example of what will happen if the alternative retrieval for dynamic gain constants leads to constants with a time difference greater than 3 days.

This is one of the failed JF runs at MID. As shown, no fixed constants retrieved, so a dynamic constants were retrieved instead and the time difference validated

Types of changes

• New feature (non-breaking change which adds functionality)

Checklist:

Reviewers

@schmidtp @mramilli @kluyvert

added Testing label

assigned to @ahmedk

added 1 commit

• 3bd65bd4 - fix warning

Compare with previous version

changed title from Account for missing fixed gain constants when burst mode in dynamic gain to [Jungfrau][Correct] Account for missing fixed gain constants when burst mode in dynamic gain

added 1 commit

• d0ae7666 - return jf_cal as well

Compare with previous version

changed the description

added 1 commit

• 6114839f - da_to_pdu definition

Compare with previous version

changed the description

Thanks! I was actually thinking of erroring out initially, but after second thought, a warning is sufficient. If there's an incident about data quality, the warning will be helpful.

changed milestone to %3.11.5

added Waiting for review label and removed Testing label

marked the checklist item TBD as completed

changed the description

    "        if abs(time_difference.days) > 3:\n",
    "            warning(\n",
    "                f\"No dynamic gain constant retrieved for {mod} with at least\"\n",
    "                \" 3 days time difference with the RAW data creation date.\"\n",
    "                \" Please make sure there are available constants.\")\n",
    "            jf_metadata[mod].pop(\"Offset10Hz\")\n",

Originally I envisioned we would raise an exception in this case. But having a warning seems a reasonable, and sufficiently document what happened in case it becomes an incident.

The CalibrationData API (v2) could probably provide something like that. Especially with trying to make aggregators less prominent, such a mapping seems useful.

(More a train of thought, ont meant as MR comment)

Thanks, LGTM

Thank you for the review!

merged

mentioned in commit dee3590a