feat: API changes for listing logs with file / level based filter sup…

…port (#1083)

* v2 API for listing logs with file / level based filter support

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Commented code removed

* Minor pre-commit fix and markdown text changes

* refactor: Removed v2 API support for execution logs, using filters instead

* Fixed org middleware behaviour for whitelisted paths

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Rahul Johny <[email protected]>

.pre-commit-config.yaml

@@ -80,7 +80,7 @@ repos:
             --settings-path=pyproject.toml,
           ]
   - repo: https://github.com/hadialqattan/pycln
-    rev: v2.4.0
+    rev: v2.5.0
     hooks:
       - id: pycln
         args: [--config=pyproject.toml]

backend/backend/public_urls_v2.py

@@ -20,9 +20,6 @@
 from django.conf.urls.static import static
 from django.urls import include, path
 
-path_prefix = settings.PATH_PREFIX
-api_path_prefix = settings.API_DEPLOYMENT_PATH_PREFIX
-
 urlpatterns = [
     path("", include("account_v2.urls")),
     # Connector OAuth

backend/backend/settings/base.py

@@ -323,7 +323,7 @@ def get_required_setting(
     "middleware.cache_control.CacheControlMiddleware",
 ]
 
-TENANT_SUBFOLDER_PREFIX = f"/{PATH_PREFIX}/unstract"
+TENANT_SUBFOLDER_PREFIX = f"{PATH_PREFIX}/unstract"
 SHOW_PUBLIC_IF_NO_TENANT_FOUND = True
 
 TEMPLATES = [
@@ -435,6 +435,11 @@ def get_required_setting(
         "django_filters.rest_framework.DjangoFilterBackend",
         "rest_framework.filters.OrderingFilter",
     ],
+    # For API versioning
+    "DEFAULT_VERSIONING_CLASS": "rest_framework.versioning.URLPathVersioning",
+    "DEFAULT_VERSION": "v1",
+    "ALLOWED_VERSIONS": ["v1"],
+    "VERSION_PARAM": "version",
 }
 
 # These paths will work without authentication

backend/file_management/constants.py

@@ -6,9 +6,3 @@ class FileInformationKey:
     FILE_UPLOAD_MAX_SIZE = 100 * 1024 * 1024
     FILE_UPLOAD_ALLOWED_EXT = ["pdf"]
     FILE_UPLOAD_ALLOWED_MIME = ["application/pdf"]
-
-
-class FileViewTypes:
-    ORIGINAL = "ORIGINAL"
-    EXTRACT = "EXTRACT"
-    SUMMARIZE = "SUMMARIZE"

backend/middleware/organization_middleware.py

@@ -6,8 +6,7 @@
 
 class OrganizationMiddleware(MiddlewareMixin):
     def process_request(self, request):
-        tenant_prefix = settings.TENANT_SUBFOLDER_PREFIX.rstrip("/") + "/"
-        pattern = rf"^{tenant_prefix}(?P<org_id>[^/]+)/"
+        pattern = r"^/api/(?P<version>v[12])/unstract/(?P<org_id>[^/]+)/"
 
         # Check if the URL matches the pattern with organization ID
         match = re.match(pattern, request.path)
@@ -17,12 +16,12 @@ def process_request(self, request):
                 re.match(path, request.path)
                 for path in settings.ORGANIZATION_MIDDLEWARE_WHITELISTED_PATHS
             ):
-                request.path_info = "/" + request.path_info
                 return
 
             org_id = match.group("org_id")
+            version = match.group("version")
             request.organization_id = org_id
-            new_path = re.sub(pattern, "/" + tenant_prefix, request.path_info)
+            new_path = re.sub(pattern, f"/api/{version}/unstract/", request.path_info)
             request.path_info = new_path
         else:
             request.organization_id = None

backend/workflow_manager/endpoint_v2/source.py

@@ -244,7 +244,7 @@ def publish_input_output_list_file_logs(
             return None
 
         folders_list = "\n".join(f"- `{folder.strip()}`" for folder in folders)
-        input_log = f"##Folders to process:\n\n{folders_list}\n\n"
+        input_log = f"## Folders to process:\n\n{folders_list}\n\n"
         self.execution_service.publish_update_log(
             state=LogState.INPUT_UPDATE, message=input_log
         )
@@ -256,9 +256,10 @@ def publish_input_output_list_file_logs(
     def publish_input_file_content(self, input_file_path: str, input_text: str) -> None:
         if not self.execution_service:
             return None
-        output_log_message = f"##Input text:\n\n```text\n{input_text}\n```\n\n"
+        output_log_message = f"## Input text:\n\n```text\n{input_text}\n```\n\n"
         input_log_message = (
-            "##Input file:\n\n```text\n" f"{os.path.basename(input_file_path)}\n```\n\n"
+            "## Input file:\n\n```text\n"
+            f"{os.path.basename(input_file_path)}\n```\n\n"
         )
         self.execution_service.publish_update_log(
             state=LogState.INPUT_UPDATE, message=input_log_message

backend/workflow_manager/urls.py

@@ -1,6 +1,6 @@
 from django.urls import include, path
 from workflow_manager.endpoint_v2 import urls as endpoint_urls
-from workflow_manager.workflow_v2 import urls as workflow_urls
+from workflow_manager.workflow_v2.urls import workflow as workflow_urls
 
 urlpatterns = [
     path("endpoint/", include(endpoint_urls)),

backend/workflow_manager/workflow_v2/execution_log_view.py

@@ -1,5 +1,6 @@
 import logging
 
+from django.db.models.query import QuerySet
 from permissions.permission import IsOwner
 from rest_framework import viewsets
 from rest_framework.versioning import URLPathVersioning
@@ -16,12 +17,24 @@ class WorkflowExecutionLogViewSet(viewsets.ModelViewSet):
     serializer_class = WorkflowExecutionLogSerializer
     pagination_class = CustomPagination
 
-    EVENT_TIME_FELID_ASC = "event_time"
+    EVENT_TIME_FIELD_ASC = "event_time"
 
-    def get_queryset(self):
+    def get_queryset(self) -> QuerySet:
         # Get the execution_id:pk from the URL path
         execution_id = self.kwargs.get("pk")
-        queryset = ExecutionLog.objects.filter(execution_id=execution_id).order_by(
-            self.EVENT_TIME_FELID_ASC
+        filter_param = {"execution_id": execution_id}
+
+        file_execution_id = self.request.query_params.get("file_execution_id")
+        if file_execution_id and file_execution_id == "null":
+            filter_param["file_execution_id"] = None
+        elif file_execution_id:
+            filter_param["file_execution_id"] = file_execution_id
+
+        log_level = self.request.query_params.get("log_level")
+        if log_level:
+            filter_param["data__level"] = log_level.upper()
+
+        queryset = ExecutionLog.objects.filter(**filter_param).order_by(
+            self.EVENT_TIME_FIELD_ASC
         )
         return queryset