Merge pull request #295 from smart-on-fhir/mikix/nlp-regression

ci: add full stack NLP regression test
smart-on-fhir · Jan 12, 2024 · 0d0db81 · 0d0db81
2 parents 15b3f51 + b143b5d
commit 0d0db81
Show file tree

Hide file tree

Showing 9 changed files with 94 additions and 9 deletions.
diff --git a/.github/workflows/ci.yaml b/.github/workflows/ci.yaml
@@ -23,7 +23,7 @@ jobs:
         python-version: ["3.10"]
 
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
 
       - name: Set up Python ${{ matrix.python-version }}
         uses: actions/setup-python@v4
@@ -37,7 +37,7 @@ jobs:
           pip install .[tests]
 
       - name: Check out MS tool
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           repository: microsoft/Tools-for-Health-Data-Anonymization
           path: mstool
@@ -57,10 +57,51 @@ jobs:
         run: |
           python -m pytest
 
+  nlp-regression:
+    runs-on: ubuntu-latest
+    env:
+      UMLS_API_KEY: ${{ secrets.UMLS_API_KEY }}
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install Docker
+        uses: docker/setup-buildx-action@v3
+
+      - name: Build ETL image
+        uses: docker/build-push-action@v5
+        with:
+          load: true  # just build, no push
+          tags: smartonfhir/cumulus-etl:latest
+
+      - name: Download NLP images
+        run: docker compose --profile covid-symptom up -d --quiet-pull
+
+      - name: Run NLP
+        run: |
+          export DATADIR=$(realpath tests/data/nlp-regression)
+
+          # Run the NLP task
+          docker compose run --rm \
+            --volume $DATADIR:/in \
+            cumulus-etl \
+            /in/input \
+            /in/run-output \
+            /in/phi \
+            --output-format=ndjson \
+            --task covid_symptom__nlp_results
+
+          # Compare results
+          export OUTDIR=$DATADIR/run-output/covid_symptom__nlp_results
+          sudo chown -R $(id -u) $OUTDIR
+          sed -i 's/"generated_on": "[^"]*", //g' $OUTDIR/*.ndjson
+          diff -upr $DATADIR/expected-output $OUTDIR
+
+          echo "All Good!"
+
   lint:
     runs-on: ubuntu-22.04
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
 
       - name: Install linters
         # black is synced with the .pre-commit-hooks version

diff --git a/.github/workflows/docker-hub.yaml b/.github/workflows/docker-hub.yaml
@@ -13,23 +13,23 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v2
+        uses: docker/setup-buildx-action@v3
 
       - name: Get Docker metadata
         id: meta
-        uses: docker/metadata-action@v4
+        uses: docker/metadata-action@v5
         with:
           flavor: latest=true
           images: smartonfhir/cumulus-etl
 
       - name: Log in to Docker Hub
-        uses: docker/login-action@v2
+        uses: docker/login-action@v3
         with:
           username: ${{ secrets.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_TOKEN }}
 
       - name: Build and push image to Docker Hub
-        uses: docker/build-push-action@v4
+        uses: docker/build-push-action@v5
         with:
           push: true
           platforms: |

diff --git a/compose.yaml b/compose.yaml
@@ -39,7 +39,7 @@ services:
   ctakes-covid-base:
     image: smartonfhir/ctakes-covid:1.1.0
     environment:
-      - ctakes_umlsuser=umls_api_key 
+      - ctakes_umlsuser=umls_api_key
       - ctakes_umlspw=$UMLS_API_KEY
     networks:
       - cumulus-etl

diff --git a/cumulus_etl/etl/studies/covid_symptom/covid_ctakes.py b/cumulus_etl/etl/studies/covid_symptom/covid_ctakes.py
@@ -73,7 +73,12 @@ async def covid_symptoms_extract(
     def is_covid_match(m: ctakesclient.typesystem.MatchText):
         return bool(covid_symptom_cuis.intersection({attr.cui for attr in m.conceptAttributes}))
 
-    matches = list(filter(is_covid_match, matches))
+    matches = filter(is_covid_match, matches)
+
+    # For better reliability when regression/unit testing, sort matches by begin / first code.
+    # (With stable sorting, we want the primary sort to be done last.)
+    matches = sorted(matches, key=lambda x: x.conceptAttributes and x.conceptAttributes[0].code)
+    matches = sorted(matches, key=lambda x: x.begin)
 
     # OK we have cTAKES symptoms. But let's also filter through cNLP transformers to remove any that are negated
     # there too. We have found this to yield better results than cTAKES alone.

diff --git a/tests/data/nlp-regression/expected-output/covid_symptom__nlp_results.000.meta b/tests/data/nlp-regression/expected-output/covid_symptom__nlp_results.000.meta
@@ -0,0 +1,9 @@
+{
+  "groups": [
+    "032b2ff6af8c883760d5a44e32ff80454d69551de6438c46be64604ddc744156",
+    "05d0686aec0a65069a1e5b1a4937f5196b75ae336b7fbe10300882184523f95e",
+    "13e748c21a7c50f6c59fc4613683cd5d7f76bd5d68fda20f4e81ccce74ea7930",
+    "364aa545eca0a9744bc67c5ad914e2e9e35dd39a5c1f1a8f902e533a8641238d",
+    "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8"
+  ]
+}
diff --git a/tests/data/nlp-regression/expected-output/covid_symptom__nlp_results.000.ndjson b/tests/data/nlp-regression/expected-output/covid_symptom__nlp_results.000.ndjson
@@ -0,0 +1,16 @@
+{"id": "032b2ff6af8c883760d5a44e32ff80454d69551de6438c46be64604ddc744156.0", "docref_id": "032b2ff6af8c883760d5a44e32ff80454d69551de6438c46be64604ddc744156", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 608, "end": 615, "text": "fatigue", "polarity": 0, "conceptAttributes": [{"code": "248274002", "cui": "C0015672", "codingScheme": "SNOMEDCT_US", "tui": "T184"}, {"code": "84229001", "cui": "C0015672", "codingScheme": "SNOMEDCT_US", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "032b2ff6af8c883760d5a44e32ff80454d69551de6438c46be64604ddc744156.1", "docref_id": "032b2ff6af8c883760d5a44e32ff80454d69551de6438c46be64604ddc744156", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 608, "end": 615, "text": "fatigue", "polarity": 0, "conceptAttributes": [{"code": "n/a", "cui": "C0015672", "codingScheme": "custom", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "032b2ff6af8c883760d5a44e32ff80454d69551de6438c46be64604ddc744156.2", "docref_id": "032b2ff6af8c883760d5a44e32ff80454d69551de6438c46be64604ddc744156", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 812, "end": 821, "text": "headaches", "polarity": 0, "conceptAttributes": [{"code": "n/a", "cui": "C0018681", "codingScheme": "custom", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "05d0686aec0a65069a1e5b1a4937f5196b75ae336b7fbe10300882184523f95e.0", "docref_id": "05d0686aec0a65069a1e5b1a4937f5196b75ae336b7fbe10300882184523f95e", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 6, "end": 14, "text": "Headache", "polarity": 0, "conceptAttributes": [{"code": "25064002", "cui": "C0018681", "codingScheme": "SNOMEDCT_US", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "05d0686aec0a65069a1e5b1a4937f5196b75ae336b7fbe10300882184523f95e.1", "docref_id": "05d0686aec0a65069a1e5b1a4937f5196b75ae336b7fbe10300882184523f95e", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 6, "end": 14, "text": "Headache", "polarity": 0, "conceptAttributes": [{"code": "n/a", "cui": "C0018681", "codingScheme": "custom", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "05d0686aec0a65069a1e5b1a4937f5196b75ae336b7fbe10300882184523f95e.2", "docref_id": "05d0686aec0a65069a1e5b1a4937f5196b75ae336b7fbe10300882184523f95e", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 114, "end": 133, "text": "nausea and vomiting", "polarity": 0, "conceptAttributes": [{"code": "16932000", "cui": "C0027498", "codingScheme": "SNOMEDCT_US", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "05d0686aec0a65069a1e5b1a4937f5196b75ae336b7fbe10300882184523f95e.3", "docref_id": "05d0686aec0a65069a1e5b1a4937f5196b75ae336b7fbe10300882184523f95e", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 114, "end": 133, "text": "nausea and vomiting", "polarity": 0, "conceptAttributes": [{"code": "n/a", "cui": "C0027498", "codingScheme": "custom", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "05d0686aec0a65069a1e5b1a4937f5196b75ae336b7fbe10300882184523f95e.4", "docref_id": "05d0686aec0a65069a1e5b1a4937f5196b75ae336b7fbe10300882184523f95e", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 603, "end": 611, "text": "fatigued", "polarity": 0, "conceptAttributes": [{"code": "n/a", "cui": "C0015672", "codingScheme": "custom", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "13e748c21a7c50f6c59fc4613683cd5d7f76bd5d68fda20f4e81ccce74ea7930.2", "docref_id": "13e748c21a7c50f6c59fc4613683cd5d7f76bd5d68fda20f4e81ccce74ea7930", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 303, "end": 318, "text": "short of breath", "polarity": 0, "conceptAttributes": [{"code": "n/a", "cui": "C0013404", "codingScheme": "custom", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8.0", "docref_id": "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 343, "end": 348, "text": "cough", "polarity": 0, "conceptAttributes": [{"code": "263731006", "cui": "C0010200", "codingScheme": "SNOMEDCT_US", "tui": "T184"}, {"code": "49727002", "cui": "C0010200", "codingScheme": "SNOMEDCT_US", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8.1", "docref_id": "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 343, "end": 348, "text": "cough", "polarity": 0, "conceptAttributes": [{"code": "n/a", "cui": "C0010200", "codingScheme": "custom", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8.2", "docref_id": "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 350, "end": 356, "text": "fevers", "polarity": 0, "conceptAttributes": [{"code": "n/a", "cui": "C0015967", "codingScheme": "custom", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8.3", "docref_id": "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 372, "end": 378, "text": "chills", "polarity": 0, "conceptAttributes": [{"code": "n/a", "cui": "C0085593", "codingScheme": "custom", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8.4", "docref_id": "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 1536, "end": 1541, "text": "fever", "polarity": 0, "conceptAttributes": [{"code": "386661006", "cui": "C0015967", "codingScheme": "SNOMEDCT_US", "tui": "T184"}, {"code": "50177009", "cui": "C0015967", "codingScheme": "SNOMEDCT_US", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8.5", "docref_id": "36ecd07bc327bba4e5ea36e34e66ca7f4f54360aef5bbcafc745c9f144aa87f8", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4, "match": {"begin": 1536, "end": 1541, "text": "fever", "polarity": 0, "conceptAttributes": [{"code": "n/a", "cui": "C0015967", "codingScheme": "custom", "tui": "T184"}], "type": "SignSymptomMention"}}
+{"id": "364aa545eca0a9744bc67c5ad914e2e9e35dd39a5c1f1a8f902e533a8641238d.0", "docref_id": "364aa545eca0a9744bc67c5ad914e2e9e35dd39a5c1f1a8f902e533a8641238d", "encounter_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "subject_id": "827db3458e3d956437c2b43f441eca441851c2f2e937e2c5467fdd0c5f980db5", "task_version": 4}