Data Eng AI

cookbook/examples/data_eng/README.md

@@ -1,44 +1,38 @@
-# Data Engineering AI
+# Data Engineering Examples
 
 > Note: Fork and clone this repository if needed
 
-1. Create a virtual environment
+### Create a virtual environment
 
 ```shell
 python3 -m venv ~/.venvs/aienv
 source ~/.venvs/aienv/bin/activate
 ```
 
-2. Install libraries
+### 2. Install libraries
 
 ```shell
 pip install -r cookbook/examples/data_eng/requirements.txt
 ```
 
-3. Run Python Assistant
+### 3. Run DuckDb Assistant
 
 ```shell
-python cookbook/examples/data_eng/python_assistant.py
-```
-
-Ask:
-
-```text
-What is the average rating of movies?
+python cookbook/examples/data_eng/duckdb_assistant.py
 ```
 
-4. Run DuckDb Assistant
+### 4. Run Python Assistant
 
 ```shell
-python cookbook/examples/data_eng/duckdb_assistant.py
+python cookbook/examples/data_eng/python_assistant.py
 ```
 
 Ask:
 
 ```text
-What is the average rating of movies? Show me the SQL.
+What is the average rating of movies?
 ```
 
-5. Message me on [discord](https://discord.gg/4MtYHHrgA8) if you have any questions
+### 5. Message me on [discord](https://discord.gg/4MtYHHrgA8) if you have any questions
 
-6. Star ⭐️ the project if you like it.
+### 6. Star ⭐️ the project if you like it.

cookbook/examples/data_eng/duckdb_assistant.py

@@ -14,8 +14,12 @@
             ]
         }
     ),
+    markdown=True,
     show_tool_calls=True,
     base_dir=Path(__file__).parent.joinpath("scratch"),
 )
 
-duckdb_assistant.cli_app(markdown=True)
+# duckdb_assistant.cli_app()
+duckdb_assistant.print_response("What is the average rating of movies? Show me the SQL?")
+duckdb_assistant.print_response("Show me a histogram of movie ratings?")
+duckdb_assistant.print_response("What are the top 5 movies?")

cookbook/examples/data_eng/requirements.txt

@@ -2,9 +2,9 @@
 # This file is autogenerated by pip-compile with Python 3.11
 # by the following command:
 #
-#    pip-compile --strip-extras cookbook/examples/data_eng/requirements.in
+#    pip-compile cookbook/examples/data_eng/requirements.in
 #
-altair==5.2.0
+altair==5.3.0
     # via streamlit
 annotated-types==0.6.0
     # via pydantic
@@ -18,14 +18,7 @@ attrs==23.2.0
     #   referencing
 blinker==1.7.0
     # via streamlit
-boto3==1.34.45
-    # via phidata
-botocore==1.34.45
-    # via
-    #   boto3
-    #   phidata
-    #   s3transfer
-cachetools==5.3.2
+cachetools==5.3.3
     # via streamlit
 certifi==2024.2.2
     # via
@@ -40,39 +33,31 @@ click==8.1.7
     #   typer
 distro==1.9.0
     # via openai
-docker==7.0.0
-    # via phidata
-duckdb==0.10.0
+duckdb==0.10.2
     # via -r cookbook/examples/data_eng/requirements.in
 gitdb==4.0.11
     # via gitpython
-gitpython==3.1.42
+gitpython==3.1.43
     # via
     #   phidata
     #   streamlit
 h11==0.14.0
     # via httpcore
-httpcore==1.0.3
+httpcore==1.0.5
     # via httpx
-httpx==0.26.0
+httpx==0.27.0
     # via
     #   openai
     #   phidata
-idna==3.6
+idna==3.7
     # via
     #   anyio
     #   httpx
     #   requests
-importlib-metadata==7.0.1
-    # via streamlit
 jinja2==3.1.3
     # via
     #   altair
     #   pydeck
-jmespath==1.0.1
-    # via
-    #   boto3
-    #   botocore
 jsonschema==4.21.1
     # via altair
 jsonschema-specifications==2023.12.1
@@ -90,44 +75,40 @@ numpy==1.26.4
     #   pyarrow
     #   pydeck
     #   streamlit
-openai==1.12.0
+openai==1.23.6
     # via -r cookbook/examples/data_eng/requirements.in
-packaging==23.2
+packaging==24.0
     # via
     #   altair
-    #   docker
     #   streamlit
-pandas==2.2.0
+pandas==2.2.2
     # via
     #   -r cookbook/examples/data_eng/requirements.in
     #   altair
     #   streamlit
 phidata==2.3.83
     # via -r cookbook/examples/data_eng/requirements.in
-pillow==10.2.0
+pillow==10.3.0
     # via streamlit
 protobuf==4.25.3
     # via streamlit
-pyarrow==15.0.0
+pyarrow==16.0.0
     # via streamlit
-pydantic==2.6.1
+pydantic==2.7.1
     # via
     #   openai
     #   phidata
     #   pydantic-settings
-pydantic-core==2.16.2
+pydantic-core==2.18.2
     # via pydantic
 pydantic-settings==2.2.1
     # via phidata
-pydeck==0.8.1b0
+pydeck==0.9.0b0
     # via streamlit
 pygments==2.17.2
     # via rich
-python-dateutil==2.8.2
-    # via
-    #   botocore
-    #   pandas
-    #   streamlit
+python-dateutil==2.9.0.post0
+    # via pandas
 python-dotenv==1.0.1
     # via
     #   phidata
@@ -136,36 +117,35 @@ pytz==2024.1
     # via pandas
 pyyaml==6.0.1
     # via phidata
-referencing==0.33.0
+referencing==0.35.0
     # via
     #   jsonschema
     #   jsonschema-specifications
 requests==2.31.0
-    # via
-    #   docker
-    #   streamlit
-rich==13.7.0
+    # via streamlit
+rich==13.7.1
     # via
     #   phidata
     #   streamlit
+    #   typer
 rpds-py==0.18.0
     # via
     #   jsonschema
     #   referencing
-s3transfer==0.10.0
-    # via boto3
+shellingham==1.5.4
+    # via typer
 six==1.16.0
     # via python-dateutil
 smmap==5.0.1
     # via gitdb
-sniffio==1.3.0
+sniffio==1.3.1
     # via
     #   anyio
     #   httpx
     #   openai
-sqlalchemy==2.0.27
+sqlalchemy==2.0.29
     # via -r cookbook/examples/data_eng/requirements.in
-streamlit==1.31.1
+streamlit==1.33.0
     # via -r cookbook/examples/data_eng/requirements.in
 tenacity==8.2.3
     # via streamlit
@@ -179,9 +159,9 @@ tornado==6.4
     # via streamlit
 tqdm==4.66.2
     # via openai
-typer==0.9.0
+typer==0.12.3
     # via phidata
-typing-extensions==4.9.0
+typing-extensions==4.11.0
     # via
     #   openai
     #   phidata
@@ -192,14 +172,5 @@ typing-extensions==4.9.0
     #   typer
 tzdata==2024.1
     # via pandas
-tzlocal==5.2
-    # via streamlit
-urllib3==2.0.7
-    # via
-    #   botocore
-    #   docker
-    #   requests
-validators==0.22.0
-    # via streamlit
-zipp==3.17.0
-    # via importlib-metadata
+urllib3==2.2.1
+    # via requests