- Add functionality to add a default dictionary for configuration fields

- Correct entitlement processing
- Remove get_template functionality from ModelVariables, define it directly with LLM model definition in configuration file.

eveai_workers/__init__.py

@@ -4,7 +4,7 @@ from flask import Flask
 import os
 
 from common.utils.celery_utils import make_celery, init_celery
-from common.extensions import db, minio_client, template_manager, cache_manager
+from common.extensions import db, minio_client, cache_manager
 import config.logging_config as logging_config
 from config.config import get_config
 
@@ -26,6 +26,8 @@ def create_app(config_file=None):
 
     register_extensions(app)
 
+    register_cache_handlers(app)
+
     from . import processors
 
     celery = make_celery(app.name, app.config)
@@ -43,7 +45,11 @@ def register_extensions(app):
     db.init_app(app)
     minio_client.init_app(app)
     cache_manager.init_app(app)
-    template_manager.init_app(app)
+
+
+def register_cache_handlers(app):
+    from common.utils.cache.config_cache import register_config_cache_handlers
+    register_config_cache_handlers(cache_manager)
 
 
 app, celery = create_app()

eveai_workers/processors/html_processor.py

@@ -3,7 +3,7 @@ from langchain_core.output_parsers import StrOutputParser
 from langchain_core.prompts import ChatPromptTemplate
 from langchain_core.runnables import RunnablePassthrough
 from common.extensions import db, minio_client
-from common.utils.model_utils import create_language_template, get_embedding_llm
+from common.utils.model_utils import create_language_template, get_embedding_llm, get_template
 from .base_processor import BaseProcessor
 from common.utils.business_event_context import current_event
 from .processor_registry import ProcessorRegistry
@@ -81,8 +81,7 @@ class HTMLProcessor(BaseProcessor):
     def _generate_markdown_from_html(self, html_content):
         self._log(f'Generating markdown from HTML for tenant {self.tenant.id}')
 
-        llm = get_embedding_llm()
-        template = self.model_variables.get_template("html_parse")
+        template, llm = get_template("html_parse")
         parse_prompt = ChatPromptTemplate.from_template(template)
         setup = RunnablePassthrough()
         output_parser = StrOutputParser()

eveai_workers/processors/pdf_processor.py

@@ -9,7 +9,7 @@ from langchain_core.runnables import RunnablePassthrough
 
 from common.eveai_model.tracked_mistral_ocr_client import TrackedMistralOcrClient
 from common.extensions import minio_client
-from common.utils.model_utils import create_language_template, get_embedding_llm
+from common.utils.model_utils import create_language_template, get_embedding_llm, get_template
 from .base_processor import BaseProcessor
 from common.utils.business_event_context import current_event
 from .processor_registry import ProcessorRegistry
@@ -208,8 +208,7 @@ class PDFProcessor(BaseProcessor):
         return text_splitter.split_text(content)
 
     def _process_chunks_with_llm(self, chunks):
-        llm = get_embedding_llm()
-        template = self.model_variables.get_template('pdf_parse')
+        template, llm = get_template('pdf_parse')
         pdf_prompt = ChatPromptTemplate.from_template(template)
         setup = RunnablePassthrough()
         output_parser = StrOutputParser()

eveai_workers/processors/transcription_processor.py

@@ -4,7 +4,7 @@ from langchain_core.output_parsers import StrOutputParser
 from langchain_core.prompts import ChatPromptTemplate
 from langchain_core.runnables import RunnablePassthrough
 
-from common.utils.model_utils import create_language_template, get_embedding_llm
+from common.utils.model_utils import create_language_template, get_embedding_llm, get_template
 from .base_processor import BaseProcessor
 from common.utils.business_event_context import current_event
 
@@ -46,8 +46,7 @@ class TranscriptionBaseProcessor(BaseProcessor):
 
     def _process_chunks(self, chunks):
         self.log_tuning("_process_chunks", {"Nr of Chunks": len(chunks)})
-        llm = get_embedding_llm()
-        template = self.model_variables.get_template('transcript')
+        template, llm = get_template('transcript')
         language_template = create_language_template(template, self.document_version.language)
         transcript_prompt = ChatPromptTemplate.from_template(language_template)
         setup = RunnablePassthrough()

eveai_workers/tasks.py

@@ -12,13 +12,13 @@ from langchain_core.runnables import RunnablePassthrough
 from sqlalchemy import or_
 from sqlalchemy.exc import SQLAlchemyError
 
-from common.extensions import db
+from common.extensions import db, cache_manager
 from common.models.document import DocumentVersion, Embedding, Document, Processor, Catalog
 from common.models.user import Tenant
 from common.utils.celery_utils import current_celery
 from common.utils.database import Database
 from common.utils.model_utils import create_language_template, get_model_variables, get_embedding_model_and_class, \
-    get_embedding_llm
+    get_embedding_llm, get_template
 
 from common.utils.business_event import BusinessEvent
 from common.utils.business_event_context import current_event
@@ -211,8 +211,8 @@ def enrich_chunks(tenant, model_variables, document_version, title, chunks):
 
 def summarize_chunk(tenant, model_variables, document_version, chunk):
     current_event.log("Starting Summarizing Chunk")
-    llm = get_embedding_llm()
-    template = model_variables.get_template("summary")
+
+    template, llm = get_template("summary")
     language_template = create_language_template(template, document_version.language)
     summary_prompt = ChatPromptTemplate.from_template(language_template)
     setup = RunnablePassthrough()