rdiff

A native python implementation of the rdiff tool by librsync


Keywords
rdiff, librsync, native, python
Install
pip install rdiff==0.2.3

Documentation

RDIFF

This project is an implementation of the rdiff tool of librsync used for finding the diff of a file on a remote machine. (https://github.com/librsync/librsync/blob/master/doc/rdiff.md).

Note:

This tool isn't concerned with how the signature, and delta files are sent/received over the network. It's only concerned with working with these files and performing the patch.

Usage:

Assume two machines, machine A and machine B, have different versions of the same file. Machine A wants to synchronise its file with the file present on machine B.
The following example shows the usage of rdiff on Machine A

import rdiff

# create an instance of the checksum class
checksum = rdiff.signature.Checksum()

# Create an instance of the signature class, passing in the checksum object
# Optionally a blocksize can be specified, defaults to 1024 bytes
signature = rdiff.signature.Signature(checksum=checksum)

# create the signature file
# basisFilePath -> Path to the file for which signature must be generated
# sigFilePath -> Where to store the signature file
signature.createSignature(basisFilePath="path_to_file", sigFilePath="path_to_signature_file")

# This machine sends over the signature file to the remote machine
# The remote machine responds back with a delta file
# rdiff isn't concerned with how this communication over the network takes place

patcher = rdiff.patch.Patch()
# Perform a patch operation
# How the delta file is obtained from the remote machine is not a concern of rdiff
# deltaFilePath -> Path to the delta file obtained from the remote machine
# basisFilePath -> Path to the file which is to be updated
# outFilePath -> Path to store the updated file
# Note: The original file isn't modified, instead a new file is created.
patcher.patchFile(
    deltaFilePath="path_to_delta_file", basisFilePath="path_to_file", outFilePath="path_to_updated_file"
)

The following shows the usage of rdiff on machine B

import rdiff

# create an instance of the checksum class
checksum = rdiff.signature.Checksum()

# create an instance of the delta class
delta = rdiff.delta.Delta()

# The createDelatFile method is used to create a delta file, for a given file
# against a signature file obtained from a remote machine.
# inFilePath -> Path to the updated file, the file which the remote machine wants to synchronise
# deltatFilePath -> Path to store the delta file
# sigFielPath -> Path to the signature file obtained from the remote machine.
# rdiff is not concerned with how this signature file is obtained.
# blocksize -> This should be identical to the blocksize used by the remote machine to generate the
# signature file.
delta.createDeltaFile(
    inFilePath="path_to_updated_file", deltaFilePath="path_to_delta_file",
    sigFielPath="path_to_sig_file", blockSize=1024, checksum=checksum
)

# This machine now sends the delta file to the remote machine.

The following is a combined example, on a single machine. The code can be extended to two remote machines.

"""
Assume there are two machines, machine A and machine B.
Machine A and machine B both have different versions of test.txt
Machine A wants to sync its file to have the same content as the file on machine B.

Machine A creates a signature file and sends it over to machine B.
Machine B uses this signature file and generates a delta file against the signature file
and sends the delta file back to machine A.

Machine A now uses this delta file to patch its file. Thereby, synchronising its file to have
the same content as the file on machine B.

The following is an example on a single machine. The same example can be extended to two
different machines connected over a network.
"""
import rdiff

checksum = rdiff.signature.Checksum()
signature = rdiff.signature.Signature(checksum=checksum, blockSize=1024)

# Machine A making the signature file
signature.createSignature(basisFilePath="path_to_file", sigFilePath="path_to_sig_file")

delta = rdiff.delta.Delta()
# Machine B creates the delta file using the signature file generated by Machine A
delta.createDeltaFile(
    inFilePath="path_to_updated_file", deltaFilePath="path_to_delta_file",
    sigFielPath="path_to_sig_file", blockSize=1024, checksum=checksum
)

patcher = rdiff.patch.Patch()
# Machine A patches its file (creates a new version of the file located at path_to_updated_file)
# using the delta file generated by Machine B
patcher.patchFile(
    deltaFilePath="path_to_delta_file", basisFilePath="path_to_file",
    outFilePath="path_to_updated_file"
)

File Formats

Information on the format/structure of the different files involved can be found at: https://github.com/MohitPanchariya/rdiff/blob/master/file_formats.md

Further Reading

More about the rsync algorithm can be found at: https://rsync.samba.org/tech_report/tech_report.html
A detailed explanation of the rsync algorithm is present in the PhD thesis of Andrew Tridgell, specifically chapter 3 of the thesis: https://www.samba.org/~tridge/phd_thesis.pdf