From a6fa9170f2ad6942d74eaf44c22cedf0b80f0fcd Mon Sep 17 00:00:00 2001
From: Ryan Ly <rly@lbl.gov>
Date: Thu, 25 May 2023 13:56:30 -0700
Subject: [PATCH] Add CI to read first NWB file from each dandiset (#1695)

---
 .github/workflows/run_dandi_read_tests.yml | 50 +++++++++++++++++++++
 CHANGELOG.md                               |  1 +
 tests/read_dandi/__init__.py               |  0
 tests/read_dandi/test_read_dandi.py        | 52 ++++++++++++++++++++++
 4 files changed, 103 insertions(+)
 create mode 100644 .github/workflows/run_dandi_read_tests.yml
 create mode 100644 tests/read_dandi/__init__.py
 create mode 100644 tests/read_dandi/test_read_dandi.py

diff --git a/.github/workflows/run_dandi_read_tests.yml b/.github/workflows/run_dandi_read_tests.yml
new file mode 100644
index 000000000..ec8cc2e84
--- /dev/null
+++ b/.github/workflows/run_dandi_read_tests.yml
@@ -0,0 +1,50 @@
+name: Run DANDI read tests
+on:
+  schedule:
+    - cron: '0 6 * * *'  # once per day at 1am ET
+  workflow_dispatch:
+
+jobs:
+  run-tests:
+    runs-on: ubuntu-latest
+    defaults:
+      run:
+        shell: bash -l {0}  # necessary for conda
+    steps:
+      - name: Cancel non-latest runs
+        uses: styfle/cancel-workflow-action@0.11.0
+        with:
+          all_but_latest: true
+          access_token: ${{ github.token }}
+
+      - uses: actions/checkout@v3
+        with:
+          submodules: 'recursive'
+          fetch-depth: 0  # tags are required for versioneer to determine the version
+
+      - name: Set up Conda
+        uses: conda-incubator/setup-miniconda@v2
+        with:
+          auto-update-conda: true
+          activate-environment: ros3
+          environment-file: environment-ros3.yml
+          python-version: "3.11"
+          channels: conda-forge
+          auto-activate-base: false
+
+      - name: Install run dependencies
+        run: |
+          python -m pip install dandi pytest
+          python -m pip uninstall -y pynwb  # uninstall pynwb
+          python -m pip install -e .
+          python -m pip list
+
+      - name: Conda reporting
+        run: |
+          conda info
+          conda config --show-sources
+          conda list --show-channel-urls
+
+      - name: Run DANDI read tests
+        run: |
+          pytest -rP tests/read_dandi/
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 542c9c6b0..24cd0985e 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -4,6 +4,7 @@
 
 ### Enhancements and minor changes
 - Add testing support for Python 3.11. @rly [#1687](https://github.com/NeurodataWithoutBorders/pynwb/pull/1687)
+- Add CI testing of NWB files on DANDI. @rly [#1695](https://github.com/NeurodataWithoutBorders/pynwb/pull/1695)
 
 ### Bug fixes
 - Remove unused, deprecated `codecov` package from dev installation requirements. @rly
diff --git a/tests/read_dandi/__init__.py b/tests/read_dandi/__init__.py
new file mode 100644
index 000000000..e69de29bb
diff --git a/tests/read_dandi/test_read_dandi.py b/tests/read_dandi/test_read_dandi.py
new file mode 100644
index 000000000..84e9f3f62
--- /dev/null
+++ b/tests/read_dandi/test_read_dandi.py
@@ -0,0 +1,52 @@
+from dandi.dandiapi import DandiAPIClient
+import sys
+import traceback
+
+from pynwb import NWBHDF5IO
+from pynwb.testing import TestCase
+
+
+class TestReadNWBDandisets(TestCase):
+    """Test reading NWB files from the DANDI Archive using ROS3."""
+
+    def test_read_first_nwb_asset(self):
+        """Test reading the first NWB asset from each dandiset that uses NWB."""
+        client = DandiAPIClient()
+        dandisets = client.get_dandisets()
+
+        failed_reads = dict()
+        for i, dandiset in enumerate(dandisets):
+            dandiset_metadata = dandiset.get_raw_metadata()
+
+            # skip any dandisets that do not use NWB
+            if not any(
+                data_standard["identifier"] == "RRID:SCR_015242"  # this is the RRID for NWB
+                for data_standard in dandiset_metadata["assetsSummary"].get("dataStandard", [])
+            ):
+                continue
+
+            dandiset_identifier = dandiset_metadata["identifier"]
+            print("--------------")
+            print(f"{i}: {dandiset_identifier}")
+
+            # iterate through assets until we get an NWB file (it could be MP4)
+            assets = dandiset.get_assets()
+            first_asset = next(assets)
+            while first_asset.path.split(".")[-1] != "nwb":
+                first_asset = next(assets)
+            if first_asset.path.split(".")[-1] != "nwb":
+                print("No NWB files?!")
+                continue
+
+            s3_url = first_asset.get_content_url(follow_redirects=1, strip_query=True)
+
+            try:
+                with NWBHDF5IO(path=s3_url, load_namespaces=True, driver="ros3") as io:
+                    io.read()
+            except Exception as e:
+                print(traceback.format_exc())
+                failed_reads[dandiset] = e
+
+        if failed_reads:
+            print(failed_reads)
+            sys.exit(1)