PyDHT is a Python implementation of distributed hashtable stored on a local machine. The implementation supports the DHT being stored in memory, on disk, or in a database. The DHT contains a list of hash values and their number of occurrences.

Follow the following steps to install PyDHT on Ubuntu:

$ git clone [email protected]:cesarghali/PyDHT.git
$ cd PyDHT
$ make
$ make test
$ sudo make install

PyDHT provides three flavors of a distributed hashtable stored on a single machine.

This class implements a DHT stored in text files on disk. Each hashtable is stored in a separate text file and includes a list of hash values and their number of occurrences in the format <hashValue:occurrences>. This class constructor expects the following parameters:

• hashSize -- the size of the hash value in bits.
• numOfHT -- the number of hashtables.
• path -- the directory path where the hashtables will be stored on disk.

One drawback of this class is that each time the number of occurrences of a specific hash value is increased, the whole corresponding hashtable file is copied.

For a usage example see LocalDiskDHTTest.

This class implements a DHT stored in memory. Each hashtable is represented as a python dictionary object. The dictionary key is the hash value itself, while the dictionary value is the number of occurrences. This class constructor expects the following parameters:

• hashSize -- the size of the hash value in bits.
• numOfHT -- the number of hashtables.

For a usage example see LocalDiskDHTTest.

This class implements a DHT stored in database on disk. Each hashtable is stored in a separate python shelve object. This class constructor expects the following parameters:

• hashSize -- the size of the hash value in bits.
• numOfHT -- the number of hashtables.
• path -- the directory path where the hashtable database files will be stored.
• syncLimit -- is the number of inserts before the corresponding shelve object is being synced.

For a usage example see LocalDiskDHTTest.

Each of the above three classes provides the following APIs.

• insert -- takes a hash value as a positive integer (or long integer) parameter and (1) inserts it in the corresponding hashtable with 1 as its number of occurrences, or (2) increases the number of occurrences by one if the hash value already exists.
• read -- takes a hash value as a positive integer (or long integer) parameter and returns its number of occurrences. If the hash value does not exist, 0 is returned.
• exists -- takes a hash value as a positive integer (or long integer) parameter and returns True if such value exists in the distributed hashtable, or False otherwise.
• calculateCollision -- calculates the probability of collisions in the current DHT.
• countCollision -- counts the hash buckets that have collision, the number of inserted hash values that collide, and the total number of hash values inserted into the DHT.

Additionally, LocalDbDHT provides a close API that closes all shelve instances.

PyDHT is licensed under the GPLv3: http://www.gnu.org/licenses/gpl-3.0.html.

PyDHT is developed and tested in Ubuntu 14.04 TLS. It should theoretically work in other environments.