Collect image dimensions from Europeana

[rwidom]

Fixes

Fixes #1484 by @stacimc

Description

Adds image dimensions, filetype, and filesize from Europeana item data.

Testing Instructions

At a high level, the goal is to run all the tests using pytest, then run the dag in airflow, then look at the data to confirm that it looks reasonable.

• Run just catalog/test-session and then pytest tests/dags/providers/provider_api_scripts/test_europeana.py.
• When I ran pytest alone from within the same test session, I got a bunch of errors from test_ingestion_server.test_index_readiness_check but that's just because the ingestion server stuff isn't started up when you're just running docker for a catalog test session.
• To get the local port for airflow (http://localhost:9090/home) working in docker, I ran just catalog/recreate. Then you can turn on the dag and it will go through daily runs and be sure to pick up some data. Check out the results using sql after just catalog/pgcli.

I temporarily replaced the dag start date with 8/8/2023 in provider_workflows.py, to get a smaller sample size for testing. Only two days out of nine had data, but it seems like the 3 second timeout might be reasonable at least for the daily runs, if these numbers of records are comparable with production. I would need to know the total number of records in production to figure out if that is workable for a refresh.

*DAG*: `europeana_workflow`
*Date range*: 2023-08-08 -> 2023-08-09
*Duration of data pull tasks*: 24 mins
*Number of records upserted per media type*:
  - `image`: 474

*DAG*: `europeana_workflow`
*Date range*: 2023-08-15 -> 2023-08-16
*Duration of data pull tasks*: 56 mins, 20 secs
*Number of records upserted per media type*:
  - `image`: 1,113

The new fields are populated in all of the inserted records:

openledger> select filetype, count(*) records, count(filesize) filesize, count(width) width, count(height) height from image group by filetype order by 
 2 desc;
+----------+---------+----------+-------+--------+
| filetype | records | filesize | width | height |
|----------+---------+----------+-------+--------|
| jpg      | 1573    | 1573     | 1573  | 1573   |
| png      | 14      | 14       | 14    | 14     |
+----------+---------+----------+-------+--------+
SELECT 2
Time: 0.020s

Checklist

• My pull request has a descriptive title (not a vague title likeUpdate index.md).
• My pull request targets the default branch of the repository (main) or a parent feature branch.
• My commit messages follow best practices.
• My code follows the established code style of the repository.
• I added or updated tests for the changes I made (if applicable).
• I added or updated documentation (if applicable).
• I tried running the project locally and verified that there are no visible errors.
• I ran the DAG documentation generator (if applicable).

Developer Certificate of Origin

Developer Certificate of Origin

Developer Certificate of Origin
Version 1.1

Copyright (C) 2004, 2006 The Linux Foundation and its contributors.
1 Letterman Drive
Suite D4700
San Francisco, CA, 94129

Everyone is permitted to copy and distribute verbatim copies of this
license document, but changing it is not allowed.


Developer's Certificate of Origin 1.1

By making a contribution to this project, I certify that:

(a) The contribution was created in whole or in part by me and I
    have the right to submit it under the open source license
    indicated in the file; or

(b) The contribution is based upon previous work that, to the best
    of my knowledge, is covered under an appropriate open source
    license and I have the right under that license to submit that
    work with modifications, whether created in whole or in part
    by me, under the same open source license (unless I am
    permitted to submit under a different license), as indicated
    in the file; or

(c) The contribution was provided directly to me by some other
    person who certified (a), (b) or (c) and I have not modified
    it.

(d) I understand and agree that this project and the contribution
    are public and that a record of the contribution (including all
    personal information I submit with it, including my sign-off) is
    maintained indefinitely and may be redistributed consistent with
    this project or the open source license(s) involved.

[obulat]

We could probably reduce nesting if we used something like
if aggregations := item_response.get("object", {}).get("aggregations"): and
for webresource in aggregation.get("webResources", [])

[obulat]

I ran the script locally and the dimensions are being picked up well. I'll defer to @stacimc on the delay timing because she has more experience with it.
I added a comment on making the code a little bit less nested, but I don't want to block on it.

[krysal]

@rwidom Thanks for adding these four fields in one PR! :) Seems like this is changing the delay of 30 seconds that we currently have. @stacimc probably has better reasons to argue about that, as Olga said, but I'll just ask: did you confirm there is no way to grab these fields from the search endpoint?

[krysal]

This looks like it could be extracted from the item_full.json file. Could it be the same for ITEM_NOT_1ST_WEBRESOURCE?

[krysal]

It would be nice to add the cases where there is only one of the dimensions.

[rwidom]

@rwidom Thanks for adding these four fields in one PR! :) Seems like this is changing the delay of 30 seconds that we currently have. @stacimc probably has better reasons to argue about that, as Olga said, but I'll just ask: did you confirm there is no way to grab these fields from the search endpoint?

I looked in the search API documentation here, and it seems like it maybe should be possible (with some kind of profile magic?), but it's not clear to me how to do it. Definitely when I look at the test files we've actually downloaded, I don't see anything that looks like those fields. It would be great if we didn't have to query individual objects, because then we could leave the 30 seconds alone.

[stacimc]

Thanks for taking this on @rwidom!

I spent some time looking at the API documentation and found the aggregations very confusing. My takeaway is that what we're getting back from the search endpoint is a collection of "metadata records", each of which may have multiple actual images associated with it. When we query the individual record endpoint, then, we may get multiple images in the aggregations with different filesizes etc.

So I took a look at one of the examples for record id /2021650/memorix_0000ee69_1b9e_6823_478f_c88af61736c6. The image url that we get from ingestion is https://images.memorix.nl/nda/thumb/fullsize/010a31d7-9316-54d7-6ad0-83b9914688d9.jpg. That url is one of the webresources we can find at the record endpoint, but it's not the first one -- that's a (smaller) version of the same image, https://images.memorix.nl/nda/thumb/640x480/010a31d7-9316-54d7-6ad0-83b9914688d9.jpg.

The result is that, at least some of the time, the final record that gets ingested will have image dimensions for a different size than the one we're linking out to, which I don't think we want. Moreover based on the API documentation, I'm worried that we might also get different photos entirely -- like related photos or multiple angles of the same object, in which case it would be really strange to get the dimensions of the wrong image.

I wonder if we could try to filter the webresources for one that exactly matches the url we used for image_url?

[rwidom]

@stacimc , I think I made the code-only changes you asked for, except the timing one. I'm not sure the best way to reach out to Europeana, but while I was looking, I came across this "Harvesting and Downloads" page which seems directly relevant to us, though I haven't looked in more detail. Hmmm...

[rwidom]

Thanks for talking through the timeout issue @stacimc and @AetherUnbound !

[stacimc]

This looks great! Just a couple tiny clean-up requests and it looks like we need to just catalog/generate-dag-docs, but then this should be good to go! 🚢

[stacimc]

🥳

[stacimc]

This all looks and tests great @rwidom! 🚀 Thanks for leading the discussion on this one, I'm excited for this :)