- Implementation of specialist execution api, including SSE protocol

- eveai_chat becomes deprecated and should be replaced with SSE
- Adaptation of STANDARD_RAG specialist
- Base class definition allowing to realise specialists with crewai framework
- Implementation of SPIN_SPECIALIST
- Implementation of test app for testing specialists (test_specialist_client). Also serves as an example for future SSE-based client
- Improvements to startup scripts to better handle and scale multiple connections
- Small improvements to the interaction forms and views
- Caching implementation improved and augmented with additional caches

eveai_api/__init__.py

@@ -15,6 +15,7 @@ from common.utils.database import Database
 from config.logging_config import LOGGING
 from .api.document_api import document_ns
 from .api.auth import auth_ns
+from .api.specialist_execution_api import specialist_execution_ns
 from config.config import get_config
 from common.utils.celery_utils import make_celery, init_celery
 from common.utils.eveai_exceptions import EveAIException
@@ -127,7 +128,7 @@ def register_extensions(app):
             "expose_headers": ["Content-Length", "Content-Range"],
             "supports_credentials": True,
             "max_age": 1728000,  # 20 days
-            "allow_credentials": True
+            # "allow_credentials": True
         }
     })
 
@@ -135,6 +136,7 @@ def register_extensions(app):
 def register_namespaces(app):
     api_rest.add_namespace(document_ns, path='/api/v1/documents')
     api_rest.add_namespace(auth_ns, path='/api/v1/auth')
+    api_rest.add_namespace(specialist_execution_ns, path='/api/v1/specialist-execution')
 
 
 def register_blueprints(app):

eveai_api/api/specialist_execution_api.py

@@ -0,0 +1,89 @@
+# eveai_api/api/specialist_execution_api.py
+import uuid
+
+from flask import Response, stream_with_context, current_app
+from flask_restx import Namespace, Resource, fields
+from flask_jwt_extended import jwt_required, get_jwt_identity
+
+from common.utils.celery_utils import current_celery
+from common.utils.execution_progress import ExecutionProgressTracker
+from eveai_api.api.auth import requires_service
+
+specialist_execution_ns = Namespace('specialist-execution', description='Specialist execution operations')
+
+specialist_start_session_response = specialist_execution_ns.model('StartSessionResponse', {
+    'session_id': fields.String(required=True, description='A new Chat session ID'),
+})
+
+
+@specialist_execution_ns.route('/start_session', methods=['GET'])
+class StartSession(Resource):
+    @jwt_required()
+    @requires_service("SPECIALIST_API")
+    @specialist_execution_ns.response(201, 'New Session ID created Successfully', specialist_start_session_response)
+    def get(self):
+        new_session_id = f"{uuid.uuid4()}"
+        return {
+            'session_id': new_session_id,
+        }, 201
+
+
+specialist_execution_input = specialist_execution_ns.model('SpecialistExecutionInput', {
+    'specialist_id': fields.Integer(required=True, description='ID of the specialist to use'),
+    'arguments': fields.Raw(required=True, description='Dynamic arguments for specialist and retrievers'),
+    'session_id': fields.String(required=True, description='Chat session ID'),
+    'user_timezone': fields.String(required=True, description='User timezone')
+})
+
+specialist_execution_response = specialist_execution_ns.model('SpecialistExecutionResponse', {
+    'task_id': fields.String(description='ID of specialist execution task, to be used to retrieve execution stream'),
+    'status': fields.String(description='Status of the execution'),
+    'stream_url': fields.String(description='Stream URL'),
+})
+
+
+@specialist_execution_ns.route('')
+class StartExecution(Resource):
+    @jwt_required()
+    @requires_service('SPECIALIST_API')
+    @specialist_execution_ns.expect(specialist_execution_input)
+    @specialist_execution_ns.response(201, 'Specialist execution successfully queued.', specialist_execution_response)
+    def post(self):
+        """Start execution of a specialist"""
+        tenant_id = get_jwt_identity()
+        data = specialist_execution_ns.payload
+
+        # Send task to queue
+        task = current_celery.send_task(
+            'execute_specialist',
+            args=[tenant_id,
+                  data['specialist_id'],
+                  data['arguments'],
+                  data['session_id'],
+                  data['user_timezone'],
+                  ],
+            queue='llm_interactions'
+        )
+
+        return {
+            'task_id': task.id,
+            'status': 'queued',
+            'stream_url': f'/api/v1/specialist-execution/{task.id}/stream'
+        }, 201
+
+
+@specialist_execution_ns.route('/<string:task_id>/stream')
+class ExecutionStream(Resource):
+    @jwt_required()
+    @requires_service('SPECIALIST_API')
+    def get(self, task_id: str):
+        """Get streaming updates for a specialist execution"""
+        progress_tracker = ExecutionProgressTracker()
+        return Response(
+            stream_with_context(progress_tracker.get_updates(task_id)),
+            mimetype='text/event-stream',
+            headers={
+                'Cache-Control': 'no-cache',
+                'Connection': 'keep-alive'
+            }
+        )