CU-8699wvj3d refactor(medcat-service): Rewrite Service - Replace Flask with FastAPI framework (#57)

* CU-8699wvj3d refactor(medcat-service): Add FastAPI framework 

* CU-8699wvj3d refactor(medcat-service): Change existing tests to hit fastapi. Add Info API

* CU-8699wvj3d refactor(medcat-service): Add retrain API though may delete later

* refactor(medcat-service): Switch to FastAPI in startup scripts

* refactor(medcat-service): Add /process API examples

* CU-8699wvj3d refactor(medcat-service): Delete Flask code

Author: alhendrickson

medcat-service/README.md

@@ -29,7 +29,7 @@ The application can be run either as a standalone Python application or as runni
 Please note that prior running the application a number of requirements need to installed (see: `requirements.txt`).
 
 There are two scripts provided implementing starting the application:
-- `start_service_debug.sh` - starts the application in the development mode and using `werkzeug` server for serving Flask applications,
+- `start_service_debug.sh` - starts the application in the development mode
 - `start_service_production.sh` - starts the application in 'production' mode and using `gunicorn` server.
 
 ## Running in a Docker container

medcat-service/medcat_service/api/__init__.py

@@ -0,0 +1,14 @@
+from functools import lru_cache
+from typing import Annotated
+
+from fastapi import Depends
+
+from medcat_service.nlp_processor.medcat_processor import MedCatProcessor
+
+
+@lru_cache
+def get_medcat_processor() -> MedCatProcessor:
+    return MedCatProcessor()
+
+
+MedCatProcessorDep = Annotated[MedCatProcessor, Depends(get_medcat_processor)]

medcat-service/medcat_service/api/api.py

@@ -0,0 +1,34 @@
+import uvicorn
+from fastapi import FastAPI, Request
+from fastapi.responses import JSONResponse
+
+from medcat_service.routers import admin, health, legacy, process
+from medcat_service.types import HealthCheckFailedException
+
+app = FastAPI(
+    title="MedCAT Service",
+    summary="MedCAT Service",
+    contact={
+        "name": "CogStack Org",
+        "url": "https://cogstack.org/",
+        "email": "[email protected]",
+    },
+    license_info={
+        "name": "Apache 2.0",
+        "identifier": "Apache-2.0",
+    },
+)
+
+app.include_router(admin.router)
+app.include_router(health.router)
+app.include_router(process.router)
+app.include_router(legacy.router)
+
+
+@app.exception_handler(HealthCheckFailedException)
+async def healthcheck_failed_exception_handler(request: Request, exc: HealthCheckFailedException):
+    return JSONResponse(status_code=503, content=exc.reason.model_dump())
+
+
+if __name__ == "__main__":
+    uvicorn.run(app, host="0.0.0.0", port=8000)

medcat-service/medcat_service/app/__init__.py

@@ -17,9 +17,10 @@
 from medcat_service.types import HealthCheckResponse, ModelCardInfo, ProcessErrorsResult, ProcessResult, ServiceInfo
 
 
-class NlpProcessor:
-    """
-    This class defines an interface for NLP Processor
+class MedCatProcessor():
+    """"
+    MedCAT Processor class is wrapper over MedCAT that implements annotations extractions functionality
+    (both single and bulk processing) that can be easily exposed for an API.
     """
 
     def __init__(self):
@@ -32,39 +33,6 @@ def __init__(self):
         self.log.debug("APP log level set to : " + str(app_log_level))
         self.log.debug("MedCAT log level set to : " + str(medcat_log_level))
 
-    def get_app_info(self):
-        pass
-
-    def process_content(self, content, *args, **kwargs):
-        pass
-
-    def process_content_bulk(self, content, *args, **kwargs):
-        pass
-
-    def is_ready(self) -> HealthCheckResponse:
-        return HealthCheckResponse(
-            name="MedCAT",
-            status="DOWN"
-        )
-
-    @staticmethod
-    def _get_timestamp():
-        """
-        Returns the current timestamp in ISO 8601 format. Formatted as "yyyy-MM-dd"T"HH:mm:ss.SSSXXX".
-        :return: timestamp string
-        """
-        return datetime.now(tz=timezone.utc).isoformat(timespec="milliseconds")
-
-
-class MedCatProcessor(NlpProcessor):
-    """"
-    MedCAT Processor class is wrapper over MedCAT that implements annotations extractions functionality
-    (both single and bulk processing) that can be easily exposed for an API.
-    """
-
-    def __init__(self):
-        super().__init__()
-
         self.log.info("Initializing MedCAT processor ...")
         self._is_ready_flag = False
 
@@ -96,6 +64,14 @@ def __init__(self):
 
         self._is_ready_flag = self._check_medcat_readiness()
 
+    @staticmethod
+    def _get_timestamp():
+        """
+        Returns the current timestamp in ISO 8601 format. Formatted as "yyyy-MM-dd"T"HH:mm:ss.SSSXXX".
+        :return: timestamp string
+        """
+        return datetime.now(tz=timezone.utc).isoformat(timespec="milliseconds")
+
     def _check_medcat_readiness(self) -> bool:
         readiness_text = "MedCAT is ready and can get_entities"
         try:
@@ -171,7 +147,7 @@ def process_content(self, content, *args, **kwargs):
             nlp_result = ProcessErrorsResult(
                 success=False,
                 errors=[error_msg],
-                timestamp=NlpProcessor._get_timestamp(),
+                timestamp=self._get_timestamp(),
             )
 
             return nlp_result
@@ -206,7 +182,7 @@ def process_content(self, content, *args, **kwargs):
             text=str(text),
             annotations=entities,
             success=True,
-            timestamp=NlpProcessor._get_timestamp(),
+            timestamp=self._get_timestamp(),
             elapsed_time=elapsed_time,
             footer=content.get("footer"),
         )
@@ -442,7 +418,7 @@ def _generate_result(self, in_documents, annotations, elapsed_time):
                     text=str(in_ct["text"]),
                     annotations=entities,
                     success=True,
-                    timestamp=NlpProcessor._get_timestamp(),
+                    timestamp=self._get_timestamp(),
                     elapsed_time=elapsed_time,
                     footer=in_ct.get("footer"),
                 )
@@ -452,7 +428,7 @@ def _generate_result(self, in_documents, annotations, elapsed_time):
                     text=str(in_ct["text"]),
                     annotations=[],
                     success=True,
-                    timestamp=NlpProcessor._get_timestamp(),
+                    timestamp=self._get_timestamp(),
                     elapsed_time=elapsed_time,
                     footer=in_ct.get("footer"),
                 )
@@ -463,7 +439,7 @@ def _generate_result(self, in_documents, annotations, elapsed_time):
                     text=str(in_ct["text"]),
                     annotations=[],
                     success=True,
-                    timestamp=NlpProcessor._get_timestamp(),
+                    timestamp=self._get_timestamp(),
                     elapsed_time=elapsed_time,
                     footer=in_ct.get("footer"),
                 )