AI-Powered PDF Analysis.
[19 Aug 2023: I've built an improved version of this tool, available here]
AI-enabled PDF Analysis
(the app might take a few seconds to wake up)
I built this PDF analysis tool for everyone. Simply upload a file and have its key insights extracted and distilled in a matter of seconds. The application utilizes OpenAI's GPT-3.5-turbo model to conduct the textual analysis.
App disclaimer: This web application is currently in beta and may producer errors. Please be aware that this application does not utilize secure file transfer protocols, and uploaded content may be accessible during transmission. Please do not upload sensitive documents. Summarizations are generated using OpenAI's large language model, and outputs may contain inaccuracies, fictional content, or potentially harmful information. Submitted information is not stored by OpenAI. For more, see OpenAI's data usage policies here. By using this tool, you acknowledge that it is provided 'as is' and without any express or implied warranties. I do not assume responsibility for any outputs. Use at your own risk and discretion. If you receive an error, would like to provide feedback, or have suggestions, please contact McKenzie.
Python3 instructions [retired version]
Update, 29th July: I have enhanced the capabilities of this tool by transitioning it from a Python3 environment to a web-based Node.js application. This shift not only boosts its functionality but also streamlines future maintenance. Please note, the original Python3 files have been retired and are no longer available.
These instructions explain how to launch a python script which uploads, analyses, and summarises a PDF file's contents, using ChatGTP.
This script was designed to summarise large PDF files, to extract the core information and present it back in a condensed manner. The script has been tested on multiple PDF files, and has been designed to minimise the number of tokens required per analysis. The program contains options regarding the strength of summarisation.
The below instructions are for Mac users, but Microsoft users should be able to follow them.
Instructions:
Download the python script & files using the below link (file: Flask PDF Analyzer)
Place the folder into an empty folder in an easy-to-locate place on your computer (e.g., documents)
Open the downloaded folder and open the file pdf_analysis.py
Find the 5th line of text which states: API_KEY = "[enter API KEY HERE]" and enter your OpenAI API Key. The line should now look something like this: API_KEY = "sijfsidfjk-gad34534534539485u34kjbsf"
Save and close the file.
Open terminal
Change directory to folder containing files and script, e.g.: cd /Users/mckenzie/flask_pdf_analyzer
Create a virtual environment by typing: python3 -m venv myenv
Note that the script requires python to run, and may not function on older versions of python. To check your python version before running the script, type: python --version or python3 --version
Activate the virtual environment by typing: source myenv/bin/activate
Ensure all repos are installed: pip3 install flask openai PyPDF2 transformers pdfminer pdfminer.six pdfplumber
Run the app: Python3 app.py
The script will run and the terminal will show debugging output. To use the application open a browser (e.g., Chrome) and enter the following URL: http://127.0.0.1:5000
You can now use the interface to upload a PDF and
Please note: This application is not regularly updated or maintained. But if you need help, please reach out via LinkedIn.