1. Make sure your device satisfies ALL THE REQUIREMENTS listed in Requirements.txt, otherwise you can't set up the environment correctly or run the script.

2. On the command line terminal, navigate to the source code folder, run 'pip install -r libraries.txt'.

3. Follow the instructions on https://pytorch.org/get-started/locally/ to install torch and torchvision. 
	PyTorch Build: Stable
	Your OS: <Your operating system>
	Package: pip
	Language: Python
	CUDA :10.1
Copy the command and run it on the terminal.

2. In order to use the script, you must have an azure account.

After creating an Azure account, you must activate Azure Cognitive Services. Go to the Cognitive Services board. Click add. Search Cognitive Services in the marketplace search bar. Click Cognitive Services. Click create. Enter the following details:
	
	Name:<Your Service Name>
	Subscription:<Your subscription plan>
	Location:<Find a location near you>
	Pricing tire:S0
	Resource group:<Your resource group>

Then, click create. After the creation, go to the Cognitive Services board. Open the Cognitive Services you just created. Click Quick start. Copy Key1 and Endpoint, and use it to replace the "subscription_key" and "endpoint" respectively.

3. Create a MySQL database on azure with "db2_full_sql.sql" database schema. Then, update the database with the credentials in "submitDataToDB.py."


4. Copy PDFs to /input folder and update pdfs table in the Azure database you created in step 3. Make sure all the PDFs under /input folder are listed in the pdfs table. Otherwise, the program won't be able to extract the information correctly.
	
	Example:
	<ID>	|<PDF_NAME>		|<NGO_NAME>
	1	|tanzania-2018.pdf	|brac
	2	|Winter-2018.pdf	|People To People Liasion PTPL

5. Run "python docQuery long context.py"