ai-parrot 0.3.5__tar.gz → 0.3.8__tar.gz

{ai_parrot-0.3.5 → ai_parrot-0.3.8}/Makefile

@@ -5,14 +5,15 @@ venv:
 install:
 	# Install Parrot
 	pip install --upgrade python-datamodel
-	pip install --upgrade asyncdb[all]
+	pip install --upgrade asyncdb[default,bigquery]
 	pip install --upgrade navconfig[default]
-	pip install --upgrade navigator-api[locale]
+	pip install --upgrade navigator-api[uvloop,locale]
 	# Nav requirements:
 	pip install --upgrade navigator-session
 	pip install --upgrade navigator-auth
 	# QS requirements
-	pip install --upgrade querysource[analytics]
+	# pip install --upgrade querysource[analytics]
+	pip install --upgrade querysource
 	# and Parrot:
 	# pip install -e .[google,milvus,groq,analytics]
 

{ai_parrot-0.3.5 → ai_parrot-0.3.8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ai-parrot
-Version: 0.3.5
+Version: 0.3.8
 Summary: Live Chatbots based on Langchain chatbots and Agents     Integrated into Navigator Framework or used into aiohttp applications.
 Home-page: https://github.com/phenobarbital/ai-parrot
 Author: Jesus Lara

{ai_parrot-0.3.5 → ai_parrot-0.3.8}/ai_parrot.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ai-parrot
-Version: 0.3.5
+Version: 0.3.8
 Summary: Live Chatbots based on Langchain chatbots and Agents     Integrated into Navigator Framework or used into aiohttp applications.
 Home-page: https://github.com/phenobarbital/ai-parrot
 Author: Jesus Lara

{ai_parrot-0.3.5 → ai_parrot-0.3.8}/ai_parrot.egg-info/SOURCES.txt

@@ -43,6 +43,7 @@ examples/extract_frames.py
 examples/load_pdf.py
 examples/test_bot.py
 examples/test_question.py
+examples/askbuddy/create_bot.py
 parrot/__init__.py
 parrot/conf.py
 parrot/exceptions.c

ai_parrot-0.3.8/examples/askbuddy/create_bot.py

@@ -0,0 +1,32 @@
+import asyncio
+from parrot.chatbots.basic import Chatbot
+from parrot.llms.vertex import VertexLLM
+
+
+async def get_agent():
+    """Return the New Agent.
+    """
+    llm = VertexLLM(
+        model='gemini-1.5-pro',
+        temperature=0.1,
+        top_k=30,
+        Top_p=0.6,
+    )
+    agent = Chatbot(
+        name='AskBuddy',
+        llm=llm
+    )
+    await agent.configure()
+    # Create the Collection
+    if agent.store.collection_exists('employee_information'):
+        await agent.store.delete_collection('employee_information')
+    await agent.store.create_collection(  # pylint: disable=E1120
+        collection_name='employee_information',
+        dimension=768,
+        index_type="IVF_SQ8",
+        metric_type='L2'
+    )
+
+
+if __name__ == "__main__":
+    agent = asyncio.run(get_agent())

{ai_parrot-0.3.5 → ai_parrot-0.3.8}/parrot/loaders/videolocal.py

@@ -1,10 +1,26 @@
 from typing import Any
 from collections.abc import Callable
+import math
 from pathlib import PurePath
 from langchain.docstore.document import Document
 from .basevideo import BaseVideoLoader
 
 
+def split_text(text, max_length):
+    """Split text into chunks of a maximum length, ensuring not to break words."""
+    chunks = []
+    while len(text) > max_length:
+        # Find the last space before the max_length
+        split_point = text.rfind(' ', 0, max_length)
+        # If no space found, split at max_length
+        if split_point == -1:
+            split_point = max_length
+        chunks.append(text[:split_point])
+        text = text[split_point:].strip()
+    chunks.append(text)
+    return chunks
+
+
 class VideoLocalLoader(BaseVideoLoader):
     """
     Generating Video transcripts from local Videos.
@@ -54,22 +70,27 @@ class VideoLocalLoader(BaseVideoLoader):
             transcript = ''
         # Summarize the transcript
         if transcript:
+            # Split transcript into chunks
+            transcript_chunks = split_text(transcript, 32767)
             summary = self.get_summary_from_text(transcript)
             # Create Two Documents, one is for transcript, second is VTT:
             metadata['summary'] = summary
-            doc = Document(
-                page_content=transcript,
-                metadata=metadata
-            )
-            documents.append(doc)
+            for chunk in transcript_chunks:
+                doc = Document(
+                    page_content=chunk,
+                    metadata=metadata
+                )
+                documents.append(doc)
         if transcript_whisper:
             # VTT version:
             transcript = self.transcript_to_vtt(transcript_whisper, transcript_path)
-            doc = Document(
-                page_content=transcript,
-                metadata=metadata
-            )
-            documents.append(doc)
+            transcript_chunks = split_text(transcript, 65535)
+            for chunk in transcript_chunks:
+                doc = Document(
+                    page_content=chunk,
+                    metadata=metadata
+                )
+                documents.append(doc)
             # Saving every dialog chunk as a separate document
             dialogs = self.transcript_to_blocks(transcript_whisper)
             docs = []

{ai_parrot-0.3.5 → ai_parrot-0.3.8}/parrot/stores/milvus.py

@@ -1,12 +1,12 @@
 from typing import Optional, Union, Any
 import asyncio
-import uuid
-import torch
+# import uuid
+# import torch
 from pymilvus import (
     MilvusClient,
-    Collection,
-    FieldSchema,
-    CollectionSchema,
+    # Collection,
+    # FieldSchema,
+    # CollectionSchema,
     DataType,
     connections,
     db
@@ -148,10 +148,10 @@ class MilvusStore(AbstractStore):
             )
 
     async def __aenter__(self):
-        try:
-            self.tensor = torch.randn(1000, 1000).cuda()
-        except RuntimeError:
-            self.tensor = None
+        # try:
+        #     self.tensor = torch.randn(1000, 1000).cuda()
+        # except RuntimeError:
+        #     self.tensor = None
         if self._embed_ is None:
             self._embed_ = self.create_embedding(
                 model_name=self.embedding_name
@@ -161,10 +161,10 @@ class MilvusStore(AbstractStore):
     async def __aexit__(self, exc_type, exc_value, traceback):
         # closing Embedding
         self._embed_ = None
-        del self.tensor
+        # del self.tensor
         try:
             self.close(alias=self._client_id)
-            torch.cuda.empty_cache()
+            # torch.cuda.empty_cache()
         except RuntimeError:
             pass
 
@@ -199,7 +199,10 @@ class MilvusStore(AbstractStore):
         connections.disconnect(alias=alias)
         try:
             self._client.close()
+        except AttributeError:
+            pass
         finally:
+            self._client = None
             self._connected = False
 
     def create_db(self, db_name: str, alias: str = 'default', **kwargs) -> bool:
@@ -269,8 +272,9 @@ class MilvusStore(AbstractStore):
         return self.get_vector()
 
     def collection_exists(self, collection_name: str) -> bool:
-        if collection_name in self._client.list_collections():
-            return True
+        with self.connection():
+            if collection_name in self._client.list_collections():
+                return True
         return False
 
     def check_state(self, collection_name: str) -> dict:
@@ -478,10 +482,10 @@ class MilvusStore(AbstractStore):
     ):
         if not collection:
             collection = self.collection
-        try:
-            tensor = torch.randn(1000, 1000).cuda()
-        except Exception:
-            tensor = None
+        # try:
+        #     tensor = torch.randn(1000, 1000).cuda()
+        # except Exception:
+        #     tensor = None
         if upsert is True:
             # get first document
             doc = documents[0]
@@ -510,7 +514,7 @@ class MilvusStore(AbstractStore):
             vector_field='vector',
             **kwargs
         )
-        del tensor
+        # del tensor
         return docstore
 
     def upsert(self, payload: dict, collection: str = None) -> None:

{ai_parrot-0.3.5 → ai_parrot-0.3.8}/parrot/version.py

@@ -3,7 +3,7 @@
 __title__ = "ai-parrot"
 __description__ = "Live Chatbots based on Langchain chatbots and Agents \
     Integrated into Navigator Framework or used into aiohttp applications."
-__version__ = "0.3.5"
+__version__ = "0.3.8"
 __author__ = "Jesus Lara"
 __author_email__ = "jesuslarag@gmail.com"
 __license__ = "MIT"