friendlydb
is a small & fast following/followers database written in Python. It can be either used directly from your Python code or over HTTP with small web API.
FriendlyDB isn't meant to be a full user system; it should be used to augment an existing system to track relationships.
Starting with v2.0.0, FriendlyDB is NOT backward-compatible with v0.4.0 & before. Prior to v2.0.0, data was stored on the filesystem, but in v2.0.0 & later, data is stored in Redis.
It was rewritten to use Redis for several reasons:
- Better performance
- Less wear/tear on hard disks
- Simpler code
However, this does mean you will need to run your own version of the Redis server (2.6.4+ recommended).
See below if you need to migrate from an older install to v2.0.0.
Using FriendlyDB from Python looks like:
from friendlydb.db import FriendlyDB
# Start using the DB (assumes Redis default host/port/db).
fdb = FriendlyDB()
# Alternatively, ``fdb = FriendlyDB(host='127.0.0.2', port=7100, db=3)``
# Grab a user by their username.
daniel = fdb['daniel']
# Follow a couple users.
daniel.follow('alice')
daniel.follow('bob')
daniel.follow('joe')
# Check the following.
daniel.following()
# Returns:
# [
# 'alice',
# 'bob',
# 'joe',
# ]
# Check joe's followers.
fdb['joe'].followers()
# Returns:
# [
# 'daniel',
# ]
# Unfollow.
daniel.unfollow('bob')
# Check the following.
daniel.following()
# Returns:
# [
# 'alice',
# 'joe',
# ]
# Dust off & nuke everything from orbit.
fdb.clear()
Using FriendlyDB from HTTP looks like (all trailing slashes are optional):
# In one shell, start the server.
python friendlydb/server.py -d /tmp/friendly
# From another, run some URLs.
curl -X GET http://127.0.0.1:8008/
# {"version": "0.3.0"}
curl -X GET http://127.0.0.1:8008/daniel/
# {"username": "daniel", "following": [], "followers": []}
curl -X POST http://127.0.0.1:8008/daniel/follow/alice/
# {"username": "daniel", "other_username": "alice", "followed": true}
curl -X POST http://127.0.0.1:8008/daniel/follow/bob/
# {"username": "daniel", "other_username": "bob", "followed": true}
curl -X POST http://127.0.0.1:8008/daniel/follow/joe/
# {"username": "daniel", "other_username": "joe", "followed": true}
curl -X POST http://127.0.0.1:8008/daniel/unfollow/joe/
# {"username": "daniel", "other_username": "joe", "unfollowed": true}
curl -X GET http://127.0.0.1:8008/daniel/
# {"username": "daniel", "following": ["alice", "bob"], "followers": []}
curl -X GET http://127.0.0.1:8008/daniel/is_following/alice/
# {"username": "daniel", "other_username": "alice", "is_following": true}
curl -X GET http://127.0.0.1:8008/alice/is_followed_by/daniel/
# {"username": "alice", "other_username": "daniel", "is_followed_by": true}
curl -X GET http://127.0.0.1:8008/alice/is_followed_by/joe/
# {"username": "alice", "other_username": "joe", "is_followed_by": false}
- Python 2.6+ or Python 3.3+
- redis.py >= 2.7.2
- (Optional) gevent for the HTTP server
- (Optional) unittest2 for running tests
Using pip, you can install it with pip install friendlydb
.
You can scope out FriendlyDB's performance for yourself by running the included benchmark.py
script.
In tests on a 2011 MacBook Pro (i7), the benchmark script demonstrated:
- created 1,000,000 relationships between 10,000 users: 179 seconds (~2.5X faster than 0.4.0)
- avg time to fetch a user's followers: 0.0016 seconds
- never exceeding 41Mb of RAM RSS
First, install & run the Redis server.
Second, run pip install redis>=2.7.2
.
To migrate your data, the easiest way is to leave your old install of FriendlyDB in place (using the HTTP server), create a new install w/ Redis, then run code like:
import requests
import json
# The new version.
from friendlydb import FriendlyDB
old_url = 'http://127.0.0.1:8008/'
fdb = FriendlyDB()
for username in users:
user = fdb[username]
# Following.
resp = requests.get("{0}/{1}/following/".format(old_url, username))
data = json.loads(resp.content)
for f_username in data.get("following", []):
user.follow(f_username)
You should create your own script & verify your data post-migration. No promises are made about the effectiveness/accuracy of the above code.
friendlydb
is maintained with passing tests at all times. Simply run:
python -m unittest2 tests
In order for a contribution to be considered for merging, it must meet the following requirements:
- Patch cleanly solves the problem
- Added test coverage (now passing) to expose the bug & check for regression
- If the behavior affects end-users, there must be docs on the changes
- The patch/tests must be compatibly licensed with New BSD
The best way to submit contributions is by forking the project on Github, applying your changes on a new branch, pushing those changes back to GH & submitting a pull request through the GitHub interface.
New BSD license.
- author
Daniel Lindsley
- version
2.0.0
- date
2013-01-17