Library for performing trends analysis via LLMs.
Project description
Trend Analyzer
Problem statement
It's time consuming to understand the changes happening in competitive / audience / creative landscapes without relying on heave pre-processing and data analysis techniques.
Solution
trend-analyzer
uses power of large language models to analyze
trends documents loaded to vector store and
make the available for querying with a simple chat bot interface.
Deliverable (implementation)
trend-analyzer
is implemented as a:
- library - Use it in your projects with a help of
TrendsAnalyzer
class. - CLI tool -
trend-analyzer
tool is available to be used in the terminal. - HTTP endpoint -
trend-analyzer
can be easily exposed as HTTP endpoint. - Langchain tool - integrated
trend-analyzer
into your Langchain applications.
Deployment
Prerequisites
-
Python 3.11+
-
A GCP project with billing account attached
-
Service account created and service account key downloaded in order to write data to BigQuery.
-
Once you downloaded service account key export it as an environmental variable
export GOOGLE_APPLICATION_CREDENTIALS=path/to/service_account.json
-
If authenticating via service account is not possible you can authenticate with the following command:
gcloud auth application-default login
-
- API key to access to access Google Gemini.
-
Once you created API key export it as an environmental variable
export GOOGLE_API_KEY=<YOUR_API_KEY_HERE>
-
Installation
Install trend-analyzer
with pip install trend-analyzer
command.
Usage
This section is focused on using
trend-analyzer
as a CLI tool. Check library, http endpoint, langchain tool sections to learn more.
Once trend-analyzer
is installed you can call it:
trend-analyzer 'YOUR_QUESTION_HERE' \
--llm gemini \
--llm.model=gemini-1.5-flash \
--output-type csv \
--output-destination sample_results
where:
YOUR_QUESTION_HERE
- trends related question you want to ask,--llm gemini
- type of large language model (currently only Google Gemini is supported)--llm.model=gemini-1.5-flash
- any parameters to initialize selected LLM--output-type csv
- type of output--output-destination sample_results
- name of output table or file.
Disclaimer
This is not an officially supported Google product.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Hashes for trend-analyzer-0.2.0.dev1.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | ea2cbcb910c85c3fc0c42d688c3ddecefc678b15f056727dff621ccd8891891c |
|
MD5 | 5733ef3a09cb52c06e3bd8d234308833 |
|
BLAKE2b-256 | d2988d381d5ee947e4095850c62583d96877089e154e27f73b86eea36dbf9770 |