Merge pull request #122 from Sunbird-AIAssistant/release-3.0.0

Merge Release 3.0.0 into main

README.md

@@ -59,15 +59,15 @@ To use the code, you need to follow these steps:
 3. To ingest data to marqo
 
     ```bash
-    python3 index_documents.py --marqo_url=<MARQO_URL> --index_name=<MARQO_INDEX_NAME> --folder_path=<PATH_TO_INPUT_FILE_DIRECTORY> --fresh_index
+    python3 index_documents.py --folder_path=<PATH_TO_INPUT_FILE_DIRECTORY> --fresh_index
     ```
    --fresh_index: This is a flag that creating a new index or overwriting an existing one. Fresh indexing typically starts from scratch without using existing data.
    PATH_TO_INPUT_FILE_DIRECTORY should have only PDF, audio, video and txt file only.
 
    e.g.
    ```bash
-   python3 index_documents.py --marqo_url=http://0.0.0.0:8882 --index_name=sakhi_parent_activities --folder_path=parent_pdfs --fresh_index
-   python3 index_documents.py --marqo_url=http://0.0.0.0:8882 --index_name=sakhi_teacher_activities --folder_path=teacher_pfs --fresh_index
+   python3 index_documents.py --folder_path=parent_pdfs --fresh_index
+   python3 index_documents.py --folder_path=teacher_pfs --fresh_index
    ```
    Create the index by using the above command. After creating the index add the index name in `config.ini` file.
 
@@ -82,32 +82,45 @@ To use the code, you need to follow these steps:
 
 5. create another file **.env** which will hold the development credentials and add the following variables. Update the Azure OpenAI details, OCI details, Bhashini endpoint URL and API key.
 
-    ```bash
-    SERVICE_ENVIRONMENT=<name_of_the_environment>
-    LOG_LEVEL=<log_level> # INFO, DEBUG, ERROR
-    CONFIG_INI_PATH=<your_config.ini_file_path>
-    OPENAI_TYPE=<openai_type> # openai, azure
-    OPENAI_API_BASE=<your_azure_openai_api_base_url>
-    OPENAI_API_VERSION=<your_openai_api_version>
-    OPENAI_API_KEY=<your_openai_api_key>
-    GPT_MODEL=<your_gpt_model>
-    TRANSLATION_TYPE=<translation_type> #bhashini, google, dhruva
-    BHASHINI_ENDPOINT_URL=<your_bhashini_api_endpoint>
-    BHASHINI_API_KEY=<your_bhashini_api_key>
-    GCP_CONFIG_PATH=<your_gcp.json_file_path>
-    BUCKET_TYPE=<bucket_type> #oci, gcp, aws
-    BUCKET_ENDPOINT_URL=<your_bucket_endpoint_url>
-    BUCKET_REGION_NAME=<your_bucket_region_name>
-    BUCKET_NAME=<your_bucket_name>  #ai-assistent-prod
-    BUCKET_SECRET_ACCESS_KEY=<your_bucket_secret_access_key>
-    BUCKET_ACCESS_KEY_ID=<your_bucket_access_key_id>
-    MARQO_URL=<your_marqo_db_url>
-    TELEMETRY_ENDPOINT_URL=<telemetry_endpoint_url>
-    TELEMETRY_LOG_ENABLED=<telemetry_enable_or_disable> # true or false
-    REDIS_HOST=localhost
-    REDIS_PORT=6379
-    REDIS_DB=0
-    ```
+   ```bash
+   SERVICE_ENVIRONMENT=<name_of_the_environment>
+   LOG_LEVEL=<log_level>
+   CONFIG_INI_PATH=<your_config.ini_file_path>
+
+   #Redis Database
+   REDIS_HOST=localhost
+   REDIS_PORT=6379
+   REDIS_DB=0
+
+   #Telemetry
+   TELEMETRY_ENDPOINT_URL=<telemetry_endpoint_url>
+   TELEMETRY_LOG_ENABLED=<telemetry_enable_or_disable>
+
+   #LLM - openai, azure, ollama
+   LLM_TYPE=<llm_type> 
+   OPENAI_API_KEY=<your_openai_api_key>
+   GPT_MODEL=<your_gpt_model>
+
+   #Translation - bhashini, google, dhruva
+   TRANSLATION_TYPE=<translation_type> 
+   BHASHINI_ENDPOINT_URL=<your_bhashini_api_endpoint>
+   BHASHINI_API_KEY=<your_bhashini_api_key>
+
+   #Storage - oci, gcp, aws
+   BUCKET_TYPE=<bucket_type>
+   BUCKET_ENDPOINT_URL=<your_bucket_endpoint_url>
+   BUCKET_REGION_NAME=<your_bucket_region_name>
+   BUCKET_NAME=<your_bucket_name>
+   BUCKET_SECRET_ACCESS_KEY=<your_bucket_secret_access_key>
+   BUCKET_ACCESS_KEY_ID=<your_bucket_access_key_id>
+
+   #Vector Store - marqo
+   VECTOR_STORE_TYPE=marqo
+   VECTOR_STORE_ENDPOINT=http://localhost:8882
+   EMBEDDING_MODEL=flax-sentence-embeddings/all_datasets_v4_mpnet-base
+   VECTOR_COLLECTION_NAME=test
+
+   ```
 
 # 🏃🏻 2. Running
 
@@ -291,7 +304,6 @@ Make the necessary changes to your dockerfile with respect to your new changes.
 | request.supported_response_format | Supported response formats                                                                     | text,audio                           |
 | request.supported_context | index name to be referred to from vector database based on context type                                                                  | teacher, parent (Default)                           |
 | llm.max_messages                   | Maximum number of messages to include in conversation history                                      |    4 |
-| llm.gpt_model                   | Gen AI GPT Model value                                                                         |                                      |
 | llm.enable_bot_intent           | Flag to enable or disable verification of user's query to check if it is referring to bot      | false                                |
 | llm.intent_prompt               | System prompt to Gen AI to verify if the user's query is referring to the bot                  |                                      |
 | llm.bot_prompt                  | System prompt to Gen AI to generate responses for user's query related to bot                  |                                      |

config.ini

@@ -16,7 +16,7 @@ supported_context = parent,teacher
 
 [llm]
 max_messages=4
-gpt_model=gpt-4
+temperature=0.3
 enable_bot_intent=false
 intent_prompt=Identify if the user's query is about the bot's persona or 'Teacher Tara' or 'Parent Tara'. If yes, return the answer as 'Yes' else return answer as 'No' only.
 bot_prompt = {
@@ -131,7 +131,6 @@ activity_prompt = {
 chat_intent_prompt=Given a chat history and the latest user question which might reference context in the chat history, formulate a standalone question which can be understood without the chat history and that can be used to find the most relevant documents. Do NOT answer the question, just reformulate it if needed and otherwise return it as is.
 [telemetry]
 telemetry_log_enabled = true
-environment = dev
 service_id = api.djp.telemetry
 service_ver = 3.1
 actor_id = sakhi-api-service

config_util.py

@@ -1,18 +1,15 @@
 import os
 from configparser import ConfigParser
-
 from fastapi import HTTPException, status
-
 from logger import logger
 
-config_file_path = os.getenv('CONFIG_INI_PATH')  # Update with your config.ini file path
+config_file_path = os.getenv('CONFIG_INI_PATH', 'config.ini')  # Update with your config.ini file path
 config = ConfigParser()
 config.read(config_file_path)
 
-
-def get_config_value(section, key, default=None):
+def get_config_value(section: str, key: str, default=None):
     # Check if the key exists in the environment variables
-    value = os.getenv(key, default)
+    value = os.getenv(key.upper(), default)
 
     # If the key is not in the environment variables, try reading from a config file
     if value is None or value == "":

env_manager.py

@@ -7,16 +7,25 @@
 from translation.utils import (
                         BhashiniTranslationClass,
                         GoogleCloudTranslationClass,
-                        DhruvaTranslationClass
+                        DhruvaTranslationClass,
+                        TranslationClass
                     )
 from storage.utils import (
                         AwsS3MainClass,
                         GoogleBucketClass,
-                        OciBucketClass
+                        OciBucketClass,
+                        StorageClass
                     )
 from llm.utils import (
-                        AzureAiClass,
-                        OpenAiClass
+                        BaseChatClient,
+                        OpenAIChatClient,
+                        AzureChatClient,
+                        OllamaChatClient
+                    )
+
+from vectorstores.utils import (
+                        MarqoVectorStore,
+                        BaseVectorStore
                     )
 
 class EnvironmentManager():
@@ -28,10 +37,11 @@ def __init__(self):
         self.indexes = {
                         "llm": {
                             "class": {
-                                "openai": OpenAiClass,
-                                "azure": AzureAiClass
+                                "openai": OpenAIChatClient,
+                                "azure": AzureChatClient,
+                                "ollama": OllamaChatClient
                             },
-                            "env_key": "OPENAI_TYPE"
+                            "env_key": "LLM_TYPE"
                         },
                         "translate": {
                             "class": {
@@ -48,21 +58,31 @@ def __init__(self):
                                 "aws": AwsS3MainClass
                             },
                             "env_key": "BUCKET_TYPE"
+                        },
+                        "vectorstore": {
+                            "class": {
+                                "marqo": MarqoVectorStore
+                            },
+                            "env_key": "VECTOR_STORE_TYPE"
                         }
                     }
 
     def create_instance(self, env_key):
         env_var = self.indexes[env_key]["env_key"]
         type_value = os.getenv(env_var)
-        logger.info(f"Init {env_key} class for: {type_value}")
-        if type_value is not None:
-            return self.indexes[env_key]["class"].get(type_value)()
 
+        if type_value is None:
+            raise ValueError(
+                f"Missing credentials. Please pass the `{env_var}` environment variable"
+            )
 
+        logger.info(f"Init {env_key} class for: {type_value}")
+        return self.indexes[env_key]["class"].get(type_value)()
+
 env_class = EnvironmentManager()
-
 # create instances of functions
 logger.info(f"Initializing required classes for components")
-ai_class = env_class.create_instance("llm")
-translate_class = env_class.create_instance("translate")
-storage_class = env_class.create_instance("storage")
+ai_class: BaseChatClient = env_class.create_instance("llm")
+translate_class: TranslationClass = env_class.create_instance("translate")
+storage_class: StorageClass = env_class.create_instance("storage")
+vectorstore_class: BaseVectorStore = env_class.create_instance("vectorstore")