nandajavarma / p2p-network

The P2P network for PeARS

p2p-network

The P2P network for PeARS

Setup PeARS by checking out to the indexer branch

Make sure you install all requirements from requirements.txt and that you have openvectors.db populated using the script uncompress_db

Build the pear profile using the code browser_history_indexer.py

python browser_history_indexer.py

Run the DHT

Clone the p2p-network code from the master branch

Start the DHT by running the following:

python dht.py 4000

This runs a standalone DHT node in your local machine in the port 4000. If you know the IP of another node that has a DHT running, then you can do the following to make the nodes identify each other:

python dht.py 4000 <Known node IP> <Known node port>

Try searching from your local history

Once the above steps are done, open up your browser and go to locahost:5000. Do try to search for something from your local index and hope for the best. :-)

Notes

The issues that need loving:

1. There is no way of telling the indexer what to index now. We end up getting a lot of pages like mail home etc. indexed which is not relevant. I have to fix this.
2. The DHT hashing is not really efficient. It randomly picks buckets, although there is a hash function that picks the node.
3. A lot of testing to do. I haven't yet tested DHT across several machines. I tested it all using docker and all the containers communicated using internal IP. I am not even sure if it works properly if DHT setup across different nodes are connected.