This pattern appears to work, executed at the root of a checkout of datasette: docker run -it -v `pwd`:/mnt --platform linux/amd64 \ datasetteproject/datasette:0.57a1 bash -c ' pip install "pytest>=5.2.2,<6.3.0" \ "pytest-xdist>=2.2.1,<2.3" \ "pytest-asyncio>=0.10,<0.16" \ "beautifulsoup4>=4.8.1,<4.10.0" \ "black==21.5b1" \ "pytest-timeout>=1.4.2,<1.5" \ "trustme>=0.7,<0.8" \ && cd /mnt && pytest'

To run the test suite... need to install a few more dependencies: https://github.com/simonw/datasette/blob/7b106e106000713bbee31b34d694b3dadbd4818c/setup.py#L71-L79

And then mount the tests/ folder into the container and run pytest.

~ % docker run -it --platform linux/arm/v7 -v `pwd`:/mnt -p 8001:8001 datasetteproject/datasette:0.57a1 python -c 'import platform; print(platform.uname())' Unable to find image 'datasetteproject/datasette:0.57a1' locally 0.57a1: Pulling from datasetteproject/datasette ee0a2cc24f29: Pull complete ca231153300d: Pull complete 6073fc53d406: Pull complete bb2a6a6421dd: Pull complete 14c15f441034: Pull complete 28d113f72ba5: Pull complete 40516c2c3785: Pull complete Digest: sha256:8dd469efa0b34b9e946c4e91496fc7bec702efec9c0f5ec548ccd4030f3263b2 Status: Downloaded newer image for datasetteproject/datasette:0.57a1 uname_result(system='Linux', node='c2e1a05732ed', release='4.19.121-linuxkit', version='#1 SMP Thu Jan 21 15:36:34 UTC 2021', machine='armv7l') Shorter version: ~ % docker run -it --platform linux/ppc64le datasetteproject/datasette:0.57a1 python -c 'import platform; print(platform.uname())' Unable to find image 'datasetteproject/datasette:0.57a1' locally 0.57a1: Pulling from datasetteproject/datasette c840eb5e9aed: Pull complete fc17d8b577ec: Pull complete 712a0ad995f8: Pull complete 56101149914c: Pull complete e6241dfe9d38: Pull complete 41ed85039888: Pull complete 26beeabe9213: Pull complete Digest: sha256:8dd469efa0b34b9e946c4e91496fc7bec702efec9c0f5ec548ccd4030f3263b2 Status: Downloaded newer image for datasetteproject/datasette:0.57a1 uname_result(system='Linux', node='0fe85887fa30', release='4.19.121-linuxkit', version='#1 SMP Thu Jan 21 15:36:34 UTC 2021', machine='ppc64le') ~ %

OK! Now I need to figure out how to run the full test suite on each one, then I can call this done.

This helps: ~ % docker run -it --platform linux/s390x -v `pwd`:/mnt \ -p 8001:8001 datasetteproject/datasette:0.57a1 \ python -c 'import platform; print(platform.uname())' uname_result(system='Linux', node='d14916ca91df', release='4.19.121-linuxkit', version='#1 SMP Thu Jan 21 15:36:34 UTC 2021', machine='s390x')

Having pushed that to Docker Hub the following seems to work on my laptop:

docker run --platform linux/s390x \
  -v `pwd`:/mnt -p 8001:8001 \
  datasetteproject/datasette:0.57a1 datasette -p 8001 -h 0.0.0.0

I'd like to confirm that this is correctly emulating the architecture and running the correct published image though.

Job completed in 13m38s - results are here: https://hub.docker.com/r/datasetteproject/datasette/tags?page=1&ordering=last_updated&name=0.57a1

Related issue: #1272 - being able to run the test suite inside Docker inside GitHub Actions would be incredibly helpful here.

https://github.com/simonw/datasette/runs/2687196350 is a run of https://github.com/simonw/datasette/blob/main/.github/workflows/push_docker_tag.yml to publish the new 0.57a1 alpha to Docker Hub.

(Should also update https://docs.datasette.io/en/stable/contributing.html#release-process with notes on how this works)

I'm going to release a 0.57a1 alpha to Docker Hub using the new buildx mechanism.

OK, since I have a mechanism for deploying alpha versions now I'm going to merge this and use it to publish 0.57a1.

That worked: % docker run -it datasetteproject/datasette:0.57a0 Unable to find image 'datasetteproject/datasette:0.57a0' locally 0.57a0: Pulling from datasetteproject/datasette 75646c2fb410: Pull complete edf8b70d3cc4: Pull complete 229daa3ebd94: Pull complete 65d21ac50a56: Pull complete e083ad7cc3ca: Pull complete ed4de1431432: Pull complete 1da5e8b27cca: Pull complete Digest: sha256:4ab61095c1d879a91d154ed469227387356a305035da9c74542c55bd500c89db Status: Downloaded newer image for datasetteproject/datasette:0.57a0 INFO: Started server process [1] INFO: Waiting for application startup. INFO: Application startup complete. INFO: Uvicorn running on http://127.0.0.1:8001 (Press CTRL+C to quit)

https://hub.docker.com/r/datasetteproject/datasette/tags?page=1&ordering=last_updated now shows this at the top: https://hub.docker.com/layers/datasetteproject/datasette/0.57a0/images/sha256-4ab61095c1d879a91d154ed469227387356a305035da9c74542c55bd500c89db?context=explore

https://github.com/simonw/datasette/actions/workflows/push_docker_tag.yml

I'm going to try pushing an alpha of https://github.com/simonw/datasette/releases/tag/0.57a0 using this workflow I built a few months ago: https://github.com/simonw/datasette/blob/1a8972f9c012cd22b088c6b70661a9c3d3847853/.github/workflows/push_docker_tag.yml

It looks like all I need to do to ship an alpha version to Docker Hub is NOT point the latest tag at it after it goes live: https://github.com/simonw/datasette/blob/1a8972f9c012cd22b088c6b70661a9c3d3847853/.github/workflows/publish.yml#L75-L77

Started a conversation about this on Twitter. Ideally I'd like to push an alpha release to Docker Hub so people can start testing these out before they make it into an official non-alpha release.

Here's a demo showing that once you get above 1,000 results you still get the "..." but it no longer links to anything: https://fivethirtyeight.datasettes.com/fivethirtyeight/births%2FUS_births_2000-2014_SSA?_facet=births&_facet_size=max

Added ?_facet_size=max in #1337 and made the ... truncated note link to that - here's a demo: https://fivethirtyeight.datasettes.com/fivethirtyeight/antiquities-act%2Factions_under_antiquities_act?_facet=states&_facet=current_agency&_facet=pres_or_congress

More exciting demo: https://fivethirtyeight.datasettes.com/fivethirtyeight/antiquities-act%2Factions_under_antiquities_act?_facet=states&_facet=current_agency&_facet=pres_or_congress

Demo: https://latest.datasette.io/fixtures/facetable?_facet_size=2&_facet=state&_facet=city_id

Turns out path_with_replaced_args() is already exposed to the table.html template context: https://github.com/simonw/datasette/blob/f1c29fd6a184254aa68efadf096bcf21e848f921/datasette/views/table.py#L897-L900

Since they are already exposed (but undocumented) this isn't actually a blocker on #1337 like I thought it was. Should still clean this up in the future though.

Turns out some of these are already being exposed to the template context on an ad-hoc basis.

In QueryView: https://github.com/simonw/datasette/blob/2bd9d54b2762c991e11950c22c88c0336158d49b/datasette/views/database.py#L295-L304

In TableView: https://github.com/simonw/datasette/blob/f1c29fd6a184254aa68efadf096bcf21e848f921/datasette/views/table.py#L897-L900

Current implementations: https://github.com/simonw/datasette/blob/51d788114035458d8f860d9ea6d74078e6c0ea0d/datasette/utils/init.py#L214-L272

In order to cleanly implement this I need to expose the path_with_replaced_args utility function to Datasette's template engine. This is the first time this will become an exposed (and hence should-by-documented) API and I don't like its shape much.

Any updates?

Here's a much easier way to implement this: implement ?_facet_size=max (sets it to max_returned_rows, default 1000) and then link to that from the ellipsis.

Codecov Report

Merging #1319 (cfca570) into main (1b69753) will increase coverage by 0.02%. The diff coverage is n/a.

```diff @@ Coverage Diff @@

main #1319 +/-

========================================== + Coverage 91.51% 91.53% +0.02%
========================================== Files 34 34
Lines 4255 4255
========================================== + Hits 3894 3895 +1
+ Misses 361 360 -1
```

| Impacted Files | Coverage Δ | | |---|---|---| | datasette/utils/__init__.py | 94.31% <0.00%> (+0.17%) | :arrow_up: |

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 1b69753...cfca570. Read the comment docs.

This is so cool! I wonder how I can best test it? I have a Raspberry Pi I should dig out to try this.

For implementation simplicity I'm going to show the item even if all of the columns are visible because ?_col= has selected all of them - selecting it will still reset the view to having no ?_col= or ?_nocol= parameters.

Demos:

• https://latest.datasette.io/fixtures/sortable?_nocol=sortable
• https://latest.datasette.io/fixtures/sortable?_col=sortable&_col=text

Also try the column cog menu on that page.

Documentation: https://docs.datasette.io/en/latest/json_api.html#special-table-arguments

Now that ?_col= and ?_nocol= are implemented I'm closing this ticket - other customizations can already be handled by defining SQL views or creating canned SQL queries.

Codecov Report

Merging #1339 (3d8598b) into main (a443dba) will increase coverage by 0.01%. The diff coverage is 91.66%.

:exclamation: Current head 3d8598b differs from pull request most recent head 387c837. Consider uploading reports for the commit 387c837 to get more accurate results

```diff @@ Coverage Diff @@

main #1339 +/-

========================================== + Coverage 91.55% 91.56% +0.01%
========================================== Files 34 34
Lines 4262 4280 +18
========================================== + Hits 3902 3919 +17
- Misses 360 361 +1
```

| Impacted Files | Coverage Δ | | |---|---|---| | datasette/views/table.py | 95.83% <90.00%> (-0.05%) | :arrow_down: | | datasette/app.py | 95.84% <100.00%> (-0.01%) | :arrow_down: | | datasette/views/database.py | 97.19% <100.00%> (ø) | |

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update a443dba...387c837. Read the comment docs.