Input: a file in .pdf format. Output: the articles from the given pdf, with metadata (text metadata, title, authors).
Because the size of the training data is over 2 GB, it will not be uploaded to repository, so you have to manually download and copy the training data to training_data folder. In order to avoid uploading by mistake the training data, the folder training_data is present in .gitignore file. You can download all the training data from here.
Download and install Node.js from here.
Download and install python 3.x from here. Tested version 3.6.4 (please use the same python version).
Download and install Tesseract OCR from here.
To install all dependencies for both client and server you can use: npm run install-all or npm run install-all-force (WARNING: The npm run install-all-force command will also edit the PATH System variables). We recommend the first option which means you'll need to edit the Environment variables by yourself adding a path for Tesseract-OCR to System variables -> Path:
C:\Program Files (x86)\Tesseract-OCRGo to root directory and use one of the following commands to run the app:
| Command | Effect |
|---|---|
npm run start |
Will start both client and server |
npm run start-newt |
Will start both client and server in new terminals |
npm run client |
Will start only the client |
npm run client-newt |
Will start only the client in a new terminal |
npm run server |
Will start only the server |
npm run server-newt |
Will start only the server in a new terminal |
You should get an output similar to this:
* Serving Flask app "doc_annotator"
* Environment: development
* Debug mode: off
* Running on http://127.0.0.1:5000/ (Press CTRL+C to quit)