Skip to content

Commit

Permalink
feat: add use case of LLM on food review running on colab (#991)
Browse files Browse the repository at this point in the history 
Co-authored-by: Jiashen Cao <jcao62@lawn-128-61-25-139.lawn.gatech.edu>
  • Loading branch information
@jiashenC
jiashenC and Jiashen Cao authored Aug 31, 2023
1 parent f31cdf4 commit 7077676
Show file tree
Hide file tree
Showing 3 changed files with 1,783 additions and 0 deletions.
2 changes: 2 additions & 0 deletions docs/_toc.yml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -18,6 +18,8 @@ parts:
title: Image Search [FAISS]
- file: source/usecases/qa-video.rst
title: Q&A from Videos [ChatGPT + HuggingFace]
- file: source/usecases/food-review.rst
title: Sentiment Analysis and Response [ChatGPT + PostgreSQL]
- file: source/usecases/02-object-detection.ipynb
title: Object Detection
- file: source/usecases/03-emotion-analysis.ipynb
Expand Down
147 changes: 147 additions & 0 deletions docs/source/usecases/food-review.rst
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,147 @@
ChatGPT + Postgres Tutorial
===========================

.. raw:: html

<embed>
<table align="left">
<td>
<a target="_blank" href="https://colab.research.google.com/github/georgia-tech-db/eva/blob/staging/tutorials/14-food-review-tone-analysis-and-response.ipynb"><img src="https://www.tensorflow.org/images/colab_logo_32px.png" /> Run on Google Colab</a>
</td>
<td>
<a target="_blank" href="https://github.com/georgia-tech-db/eva/blob/staging/tutorials/14-food-review-tone-analysis-and-response.ipynb"><img src="https://www.tensorflow.org/images/GitHub-Mark-32px.png" /> View source on GitHub</a>
</td>
<td>
<a target="_blank" href="https://github.com/georgia-tech-db/eva/raw/staging/tutorials/14-food-review-tone-analysis-and-response.ipynb"><img src="https://www.tensorflow.org/images/download_logo_32px.png" /> Download notebook</a>
</td>
</table><br><br>
</embed>

In this tutorial, we demonstrate how to use EvaDB + ChatGPT to analyze the tone of food reviews stored in PostgreSQL. Then, based on the analysis, we further use
EvaDB + ChatGPT to address negative reviews by proposing a solution to the customer.

For this use case, we assume user has a Postgres server running locally. You can also check our notebook above to skip Postgres setup.

1. Connect to EvaDB
-------------------

.. code-block:: python
import evadb
cursor = evadb.connect().cursor()
2. Connect to an Existing Postgres Database
-------------------------------------------

.. tab-set::

.. tab-item:: Python

.. code-block:: python
params = {
"user": "eva",
"password": "password",
"host": "localhost",
"port": "5432",
"database": "evadb",
}
query = f"CREATE DATABASE postgres_data WITH ENGINE = 'postgres', PARAMETERS = {params};"
cursor.query(query).df()
.. tab-item:: SQL

.. code-block:: sql
CREATE DATABASE postgres_data WITH ENGINE = 'postgres', PARAMETERS = {
"user": "eva",
"password": "password",
"host": "localhost",
"port": "5432",
"database": "evadb"
}
3. Sentiment Analysis of Food Review using ChatGPT
---------------------------------------------

We then use EvaDB + ChatGPT to analyze whether the review is "positive" or "negative" with customized ChatGPT prompt. For this use case,
we assume reviews have been already loaded into the table inside PostgreSQL.
You can check our `Jupyter Notebook <https://github.com/georgia-tech-db/eva/blob/staging/tutorials/14-food-review-tone-analysis-and-response.ipynb>`_ for how to load data.

.. tab-set::

.. tab-item:: Python

.. code-block:: python
cursor.query("""
SELECT ChatGPT(
"Is the review positive or negative. Only reply 'positive' or 'negative'. Here are examples. The food is very bad: negative. The food is very good: postive.",
review)
FROM postgres_data.review_table;
""").df()
.. tab-item:: SQL

.. code-block:: sql
SELECT ChatGPT(
"Is the review positive or negative. Only reply 'positive' or 'negative'. Here are examples. The food is very bad: negative. The food is very good: postive.",
review)
FROM postgres_data.review_table;
This will return tone analysis results for existing reviews.

.. code-block::
+------------------------------+
| chatgpt.response |
|------------------------------|
| negative |
| positive |
| negative |
+------------------------------+
4. Response to Negative Reviews using ChatGPT
---------------------------------------------

.. tab-set::

.. tab-item:: Python

.. code-block:: python
cursor.query("""
SELECT ChatGPT(
"Respond the the review with solution to address the review's concern",
review)
FROM postgres_data.review_table
WHERE ChatGPT(
"Is the review positive or negative. Only reply 'positive' or 'negative'. Here are examples. The food is very bad: negative. The food is very good: postive.",
review) = "negative";
""").df()
.. tab-item:: SQL

.. code-block:: sql
SELECT ChatGPT(
"Respond the the review with solution to address the review's concern",
review)
FROM postgres_data.review_table
WHERE ChatGPT(
"Is the review positive or negative. Only reply 'positive' or 'negative'. Here are examples. The food is very bad: negative. The food is very good: postive.",
review) = "negative";
This query will first filter out positive reviews and then apply ChatGPT again to create response to negative reviews. This will give results.

.. code-block::
+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| chatgpt.response |
|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Dear valued customer, Thank you for bringing this matter to our attention. We apologize for the inconvenience caused by the excessive saltiness of your fried rice. We understand how important it is to have a satisfying dining experience, and we would like to make it right for you ... |
| Dear [Customer's Name], Thank you for bringing this issue to our attention. We apologize for the inconvenience caused by the missing chicken sandwich in your takeout order. We understand how frustrating it can be when an item is missing from your meal. To address this concern, we ... |
+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
Check out our `Jupyter Notebook <https://github.com/georgia-tech-db/evadb/blob/staging/tutorials/14-food-review-tone-analysis-and-response.ipynb>`_ for working example.
Loading

0 comments on commit 7077676

Please # to comment.