Pathfinder scoring model

.gitignore

@@ -168,3 +168,5 @@ cython_debug/
 #  and can be added to the global gitignore or merged into this file.  For a more nuclear
 #  option (not recommended) you can uncomment the following to ignore the entire idea folder.
 #.idea/
+
+gandalf_mmap/

README.md

@@ -12,7 +12,10 @@ The main entrypoint is `./compose.yml` and will spin everything up.
 
 - In the root folder, run `docker compose up --build`
 
-If you want to add a new operation/worker, add a new service in `compose.yml` under `services`.
+If you would like to run workers with gpu access, you can use the `./compose.gpu.yml` override file. Otherwise, workers will be restricted to cpu only.
+- In the root folder, run `docker compose -f compose.yml -f compose.gpu.yml up --build`
+
+If you want to add a new operation/worker, add a new service in `compose.yml` under `services`. If you want to add a new operation/worker that can utilize a gpu, add the service without the gpu deployment into `compose.yml` and then use the override file `compose.gpu.yml` to add the gpu as a resource so it may be accessible within the container.
 
 ### Worker
 

compose.gpu.yml

@@ -0,0 +1,12 @@
+services:
+  score_paths:
+    build:
+      args:
+        DEVICE: gpu
+    deploy:
+      resources:
+        reservations:
+          devices:
+            - driver: nvidia
+              count: all
+              capabilities: [gpu]

compose.yml

@@ -194,6 +194,8 @@ services:
     build:
       context: .
       dockerfile: workers/score_paths/Dockerfile
+      args:
+        DEVICE: cpu
     restart: unless-stopped
     depends_on:
       shepherd_db:
@@ -204,7 +206,6 @@ services:
       - ./logs:/app/logs
       - ./.env:/app/.env
 
-
   ######### Example ARA
   example_ara:
     container_name: example_ara

shepherd_server/base_routes.py

@@ -144,7 +144,9 @@ async def run_sync_query(
     query_id, response_id, logger = await run_query(target, query_dict)
     start = time.time()
     now = start
-    while now <= start + 360:
+    timeout = query_dict.get("parameters", {}).get("timeout", 360)
+    logger.info(f"Query running with {timeout} second timeout.")
+    while now <= start + timeout:
         now = time.time()
         # poll for completed status
         query_state = await get_query_state(query_id, logger)

shepherd_server/openapi-config.yaml

@@ -4,7 +4,7 @@ contact:
   x-id: https://github.com/maximusunc
   x-role: responsible developer
 description: '<img src="/static/favicon.png" width="200px"><br /><br />Shepherd: Translator Autonomous Relay Agent Platform'
-version: 0.6.9
+version: 0.7.0
 servers:
 - description: Default server
   url: https://shepherd.renci.org

workers/aragorn_pathfinder/worker.py

@@ -94,11 +94,12 @@ async def shadowfax(task, logger: logging.Logger):
                         "subject": pinned_node_keys[0],
                         "object": "intermediate_0",
                         "predicates": [
+                            "biolink:physically_interacts_with",
+                            "biolink:genetically_interacts_with",
                             "biolink:contributes_to",
                             "biolink:contribution_from",
                             "biolink:affects",
                             "biolink:affected_by",
-                            "biolink:interacts_with",
                             "biolink:acts_upstream_of",
                             "biolink:has_upstream_actor",
                             "biolink:enables",
@@ -115,17 +116,19 @@ async def shadowfax(task, logger: logging.Logger):
                             "biolink:translation_of",
                             "biolink:has_gene_product",
                             "biolink:gene_product_of",
+                            "biolink:genetically_associated_with",
                         ],
                     },
                     "e1": {
                         "subject": "intermediate_0",
                         "object": "intermediate_1",
                         "predicates": [
+                            "biolink:physically_interacts_with",
+                            "biolink:genetically_interacts_with",
                             "biolink:contributes_to",
                             "biolink:contribution_from",
                             "biolink:affects",
                             "biolink:affected_by",
-                            "biolink:interacts_with",
                             "biolink:acts_upstream_of",
                             "biolink:has_upstream_actor",
                             "biolink:enables",
@@ -142,17 +145,19 @@ async def shadowfax(task, logger: logging.Logger):
                             "biolink:translation_of",
                             "biolink:has_gene_product",
                             "biolink:gene_product_of",
+                            "biolink:genetically_associated_with",
                         ],
                     },
                     "e2": {
                         "subject": "intermediate_1",
                         "object": pinned_node_keys[1],
                         "predicates": [
+                            "biolink:physically_interacts_with",
+                            "biolink:genetically_interacts_with",
                             "biolink:contributes_to",
                             "biolink:contribution_from",
                             "biolink:affects",
                             "biolink:affected_by",
-                            "biolink:interacts_with",
                             "biolink:acts_upstream_of",
                             "biolink:has_upstream_actor",
                             "biolink:enables",
@@ -169,6 +174,7 @@ async def shadowfax(task, logger: logging.Logger):
                             "biolink:translation_of",
                             "biolink:has_gene_product",
                             "biolink:gene_product_of",
+                            "biolink:genetically_associated_with",
                         ],
                     },
                 },

workers/score_paths/Dockerfile

@@ -4,27 +4,30 @@ FROM ghcr.io/translatorsri/renci-python-image:3.11.5
 # Add image info
 LABEL org.opencontainers.image.source https://github.com/BioPack-team/shepherd
 
+ARG DEVICE=gpu
+
 ENV PYTHONHASHSEED=0
+ENV HF_HOME=/app/hf_cache
 
 # set up requirements
 WORKDIR /app
 
-# make sure all is writeable for the nru USER later on
-RUN chmod -R 777 .
+COPY ./workers/score_paths/requirements-cpu.txt ./workers/score_paths/requirements-gpu.txt ./
+RUN pip install --no-cache-dir --timeout=300 --retries=5 -r requirements-${DEVICE}.txt
+# pre-download the model
+RUN python -c "from sentence_transformers import SentenceTransformer; SentenceTransformer('cambridgeltl/SapBERT-from-PubMedBERT-fulltext')"
 
-# Install requirements
+# Install shared utilities
 COPY ./shepherd_utils ./shepherd_utils
 COPY ./pyproject.toml .
 RUN pip install .
 
-COPY ./workers/score_paths/requirements.txt .
-RUN pip install -r requirements.txt
+# Copy in files and make sure they are readable
+COPY ./workers/score_paths .
+RUN chmod -R 777 .
 
 # switch to the non-root user (nru). defined in the base image
 USER nru
 
-# Copy in files
-COPY ./workers/score_paths .
-
 # Variables that can be overriden
 CMD ["python", "worker.py"]

workers/score_paths/requirements-cpu.txt

@@ -0,0 +1,5 @@
+xgboost>=3.2.0
+sentence-transformers>=5.2.3
+--extra-index-url https://download.pytorch.org/whl/cpu
+torch>=2.10.0
+bmt>=1.4.6

workers/score_paths/requirements-gpu.txt

@@ -0,0 +1,4 @@
+xgboost>=3.2.0
+sentence-transformers>=5.2.3
+torch>=2.10.0
+bmt>=1.4.6