ML/LLM: Updates for compatibility with LangChain 0.2

- Imports: `langchain-cli migrate --include-ipynb .`
- SQLDatabaseLoader superseded SQLAlchemyLoader

Author: amotl

topic/machine-learning/llm-langchain/conversational_memory.ipynb

@@ -56,7 +56,7 @@
    "execution_count": 2,
    "outputs": [],
    "source": [
-    "from langchain.memory.chat_message_histories import CrateDBChatMessageHistory\n",
+    "from langchain_community.chat_message_histories import CrateDBChatMessageHistory\n",
     "\n",
     "# Connect to a self-managed CrateDB instance.\n",
     "CONNECTION_STRING = \"crate://crate@localhost/?schema=notebook\"\n",
@@ -153,11 +153,10 @@
     "from datetime import datetime\n",
     "from typing import Any\n",
     "\n",
-    "from langchain.memory.chat_message_histories.sql import BaseMessageConverter\n",
-    "from langchain.schema import AIMessage, BaseMessage, HumanMessage, SystemMessage\n",
-    "\n",
     "import sqlalchemy as sa\n",
     "from sqlalchemy.orm import declarative_base\n",
+    "from langchain_community.chat_message_histories.sql import BaseMessageConverter\n",
+    "from langchain_core.messages import AIMessage, BaseMessage, HumanMessage, SystemMessage\n",
     "\n",
     "\n",
     "Base = declarative_base()\n",
@@ -284,7 +283,7 @@
     "import json\n",
     "import typing as t\n",
     "\n",
-    "from langchain.memory.chat_message_histories.cratedb import CrateDBMessageConverter\n",
+    "from langchain_community.chat_message_histories.cratedb import CrateDBMessageConverter\n",
     "from langchain.schema import _message_to_dict\n",
     "\n",
     "\n",

topic/machine-learning/llm-langchain/conversational_memory.py

@@ -16,7 +16,7 @@
 import os
 from pprint import pprint
 
-from langchain.memory.chat_message_histories import CrateDBChatMessageHistory
+from langchain_community.chat_message_histories import CrateDBChatMessageHistory
 
 
 CONNECTION_STRING = os.environ.get(

topic/machine-learning/llm-langchain/cratedb-vectorstore-rag-openai-sql.ipynb

@@ -79,12 +79,12 @@
     "import pandas as pd\n",
     "import sqlalchemy as sa\n",
     "\n",
-    "from langchain.document_loaders import PyPDFLoader\n",
-    "from langchain.embeddings.openai import OpenAIEmbeddings\n",
-    "from langchain.text_splitter import RecursiveCharacterTextSplitter\n",
-    "\n",
     "# TODO: Bring this into the `crate-python` driver.\n",
     "from cratedb_toolkit.sqlalchemy.patch import patch_inspector\n",
+    "from langchain_community.document_loaders import PyPDFLoader\n",
+    "from langchain_text_splitters import RecursiveCharacterTextSplitter\n",
+    "from langchain_openai import OpenAIEmbeddings\n",
+    "\n",
     "patch_inspector()"
    ]
   },

topic/machine-learning/llm-langchain/cratedb_rag_customer_support.ipynb

@@ -61,13 +61,8 @@
    },
    "outputs": [],
    "source": [
-    "from langchain.document_loaders.csv_loader import CSVLoader\n",
-    "from langchain.vectorstores import Chroma\n",
-    "from langchain.chat_models import ChatOpenAI\n",
-    "from langchain.text_splitter import RecursiveCharacterTextSplitter\n",
     "from langchain.chains import RetrievalQA, ConversationalRetrievalChain\n",
-    "from langchain.llms import OpenAI\n",
-    "from langchain_openai import OpenAIEmbeddings\n",
+    "from langchain_openai import ChatOpenAI, OpenAI, OpenAIEmbeddings\n",
     "import pandas as pd\n",
     "import sqlalchemy as sa\n",
     "from sqlalchemy import create_engine\n",
@@ -76,7 +71,10 @@
     "import openai\n",
     "import os\n",
     "import requests\n",
-    "from pueblo.util.environ import getenvpass"
+    "from pueblo.util.environ import getenvpass\n",
+    "from langchain_community.document_loaders import CSVLoader\n",
+    "from langchain_community.vectorstores import Chroma\n",
+    "from langchain_text_splitters import RecursiveCharacterTextSplitter"
    ]
   },
   {

topic/machine-learning/llm-langchain/cratedb_rag_customer_support_langchain.ipynb

@@ -105,9 +105,9 @@
     "import requests\n",
     "\n",
     "from pueblo.util.environ import getenvpass\n",
-    "from langchain.document_loaders.csv_loader import CSVLoader\n",
     "from langchain_openai import OpenAIEmbeddings\n",
-    "from langchain.vectorstores import CrateDBVectorSearch\n",
+    "from langchain_community.document_loaders import CSVLoader\n",
+    "from langchain_community.vectorstores import CrateDBVectorSearch\n",
     "\n",
     "warnings.filterwarnings('ignore')"
    ]

topic/machine-learning/llm-langchain/cratedb_rag_customer_support_vertexai.ipynb

@@ -107,8 +107,6 @@
     "\n",
     "\n",
     "from pueblo.util.environ import getenvpass\n",
-    "from langchain.embeddings import VertexAIEmbeddings\n",
-    "from langchain.llms import VertexAI\n",
     "from google.cloud import aiplatform\n",
     "from vertexai.generative_models import (\n",
     "    GenerationConfig,\n",
@@ -117,9 +115,10 @@
     "    HarmBlockThreshold,\n",
     "    HarmCategory,\n",
     ")\n",
-    "\n",
-    "from langchain.document_loaders.csv_loader import CSVLoader\n",
-    "from langchain.vectorstores import CrateDBVectorSearch\n",
+    "from langchain_community.document_loaders import CSVLoader\n",
+    "from langchain_community.embeddings import VertexAIEmbeddings\n",
+    "from langchain_community.llms import VertexAI\n",
+    "from langchain_community.vectorstores import CrateDBVectorSearch\n",
     "\n",
     "warnings.filterwarnings('ignore')"
    ]

topic/machine-learning/llm-langchain/document_loader.ipynb

@@ -103,12 +103,16 @@
    },
    "outputs": [],
    "source": [
-    "from langchain.document_loaders import CrateDBLoader\n",
+    "import sqlalchemy as sa\n",
+    "from langchain_community.document_loaders import CrateDBLoader\n",
+    "from langchain_community.utilities.sql_database import SQLDatabase\n",
     "from pprint import pprint\n",
     "\n",
+    "db = SQLDatabase(engine=sa.create_engine(CONNECTION_STRING))\n",
+    "\n",
     "loader = CrateDBLoader(\n",
     "    'SELECT * FROM mlb_teams_2012 ORDER BY \"Team\" LIMIT 5;',\n",
-    "    url=CONNECTION_STRING,\n",
+    "    db=db,\n",
     ")\n",
     "documents = loader.load()"
    ]
@@ -151,7 +155,7 @@
    "source": [
     "loader = CrateDBLoader(\n",
     "    'SELECT * FROM mlb_teams_2012 ORDER BY \"Team\" LIMIT 5;',\n",
-    "    url=CONNECTION_STRING,\n",
+    "    db=db,\n",
     "    page_content_columns=[\"Team\"],\n",
     "    metadata_columns=[\"Payroll (millions)\"],\n",
     ")\n",
@@ -194,7 +198,7 @@
    "source": [
     "loader = CrateDBLoader(\n",
     "    'SELECT * FROM mlb_teams_2012 ORDER BY \"Team\" LIMIT 5;',\n",
-    "    url=CONNECTION_STRING,\n",
+    "    db=db,\n",
     "    source_columns=[\"Team\"],\n",
     ")\n",
     "documents = loader.load()"

topic/machine-learning/llm-langchain/document_loader.py

@@ -27,8 +27,10 @@
 import os
 
 import requests
+import sqlalchemy as sa
 from cratedb_toolkit.util import DatabaseAdapter
-from langchain.document_loaders import CrateDBLoader
+from langchain_community.document_loaders import CrateDBLoader
+from langchain_community.utilities.sql_database import SQLDatabase
 from pprint import pprint
 
 
@@ -55,10 +57,12 @@ def main():
     # Load data.
     import_mlb_teams_2012()
 
+    db = SQLDatabase(engine=sa.create_engine(CONNECTION_STRING))
+
     # Query data.
     loader = CrateDBLoader(
         query="SELECT * FROM mlb_teams_2012 LIMIT 3;",
-        url=CONNECTION_STRING,
+        db=db,
         include_rownum_into_metadata=True,
     )
     docs = loader.load()

topic/machine-learning/llm-langchain/requirements.txt

@@ -9,6 +9,7 @@ cratedb-toolkit==0.0.12
 google-cloud-aiplatform<2
 langchain-google-vertexai<2
 langchain-openai<0.2
+langchain-text-splitters<0.3
 pueblo[cli,nlp]>=0.0.7
 pydantic>=1,<3
 pypdf<5

topic/machine-learning/llm-langchain/vector_search.ipynb

@@ -163,9 +163,9 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "from langchain.docstore.document import Document\n",
-    "from langchain.embeddings.openai import OpenAIEmbeddings\n",
-    "from langchain.vectorstores import CrateDBVectorSearch"
+    "from langchain_community.vectorstores import CrateDBVectorSearch\n",
+    "from langchain_core.documents import Document\n",
+    "from langchain_openai import OpenAIEmbeddings"
    ]
   },
   {

topic/machine-learning/llm-langchain/vector_search.py

@@ -19,10 +19,10 @@
     # Run program.
     python vector_search.py
 """  # noqa: E501
-from langchain.document_loaders import UnstructuredURLLoader
-from langchain.embeddings.openai import OpenAIEmbeddings
-from langchain.text_splitter import CharacterTextSplitter
-from langchain.vectorstores import CrateDBVectorSearch
+from langchain_community.document_loaders import UnstructuredURLLoader
+from langchain_community.vectorstores import CrateDBVectorSearch
+from langchain_text_splitters import CharacterTextSplitter
+from langchain_openai import OpenAIEmbeddings
 
 
 def main():