etcetra

Pure python asyncio Etcd client.

Installation

pip install etcetra

API Documentation

Refer here.

Basic usage

All etcd operations managed by etcetra can be executed using EtcdClient. EtcdClient instance is a wrapper which holds connection information to Etcd channel. This instance is reusable, since actual connection to gRPC channel will be established when you initiate connection calls (see below).

from etcetra import EtcdClient, HostPortPair
etcd = EtcdClient(HostPortPair('127.0.0.1', 2379))

Like I mentioned above, actual connection establishment with Etcd's gRPC channel will be done when you call EtcdClient.connect(). This call returns async context manager, which manages EtcdCommunicator instance.

async with etcd.connect() as communicator:
    await communicator.put('testkey', 'testvalue')
    value = await communicator.get('testkey')
    print(value)  # testvalue

EtcdCommunicator.get_prefix(prefix) will return a dictionary containing all key-values with given key prefix.

async with etcd.connect() as communicator:
    await communicator.put('/testdir', 'root')
    await communicator.put('/testdir/1', '1')
    await communicator.put('/testdir/2', '2')
    await communicator.put('/testdir/2/3', '3')
    test_dir = await communicator.get_prefix('/testdir')
    print(test_dir)  # {'/testdir': 'root', '/testdir/1': '1', '/testdir/2': '2', '/testdir/2/3': '3'}

Operating with Etcd lock

Just like EtcdClient.connect(), you can easilly use etcd lock by calling EtcdClient.with_lock(lock_name, timeout=None).

async def first():
    async with etcd.with_lock('foolock') as communicator:
        value = await communicator.get('testkey')
        print('first:', value, end=' | ')

async def second():
    await asyncio.sleep(0.1)
    async with etcd.with_lock('foolock') as communicator:
        value = await communicator.get('testkey')
        print('second:', value)

async with etcd.connect() as communicator:
    await communicator.put('testkey', 'testvalue')
await asyncio.gather(first(), second())  # first: testvalue | second: testvalue

Adding timeout parameter to EtcdClient.with_lock() call will add a timeout to lock acquiring process.

async def first():
    async with etcd.with_lock('foolock') as communicator:
        value = await communicator.get('testkey')
        print('first:', value)
        await asyncio.sleep(10)

async def second():
    await asyncio.sleep(0.1)
    async with etcd.with_lock('foolock', timeout=5) as communicator:
        value = await communicator.get('testkey')
        print('second:', value)

async with etcd.connect() as communicator:
    await communicator.put('testkey', 'testvalue')
await asyncio.gather(first(), second())  # asyncio.TimeoutError followed by first: testvalue output

Watch

You can watch changes on key with EtcdCommunicator.watch(key).

async def watch():
    async with etcd.connect() as communicator:
        async for event in communicator.watch('testkey'):
            print(event.event, event.value)

async def update():
    await asyncio.sleep(0.1)
    async with etcd.connect() as communicator:
        await communicator.put('testkey', '1')
        await communicator.put('testkey', '2')
        await communicator.put('testkey', '3')
        await communicator.put('testkey', '4')
        await communicator.put('testkey', '5')

await asyncio.gather(watch(), update())
# WatchEventType.PUT 1
# WatchEventType.PUT 2
# WatchEventType.PUT 3
# WatchEventType.PUT 4
# WatchEventType.PUT 5

Watching changes on keys with specific prefix can be also done by EtcdCommunicator.watch_prefix(key_prefix).

async def watch():
    async with etcd.connect() as communicator:
        async for event in communicator.watch_prefix('/testdir'):
            print(event.event, event.key, event.value)

async def update():
    await asyncio.sleep(0.1)
    async with etcd.connect() as communicator:
        await communicator.put('/testdir', '1')
        await communicator.put('/testdir/foo', '2')
        await communicator.put('/testdir/bar', '3')
        await communicator.put('/testdir/foo/baz', '4')

await asyncio.gather(watch(), update())
# WatchEventType.PUT /testdir 1
# WatchEventType.PUT /testdir/foo 2
# WatchEventType.PUT /testdir/bar 3
# WatchEventType.PUT /testdir/foo/baz 4

Transaction

You can run etcd transaction by calling EtcdCommunicator.txn_compare(compares, txn_builder).

Constructing compares

Constructing compare operations can be done by comparing CompareKey instance with value with Python's built-in comparison operators (==, !=, >, <).

from etcetra import CompareKey
compares = [
    CompareKey('cmpkey1').value == 'foo',
    CompareKey('cmpkey2').value > 'bar',
]

Executing transaction calls

async with etcd.connect() with communicator:
    await communicator.put('cmpkey1', 'foo')
    await communicator.put('cmpkey2', 'baz')
    await communicator.put('successkey', 'asdf')

    def _txn(success, failure):
        success.get('successkey')

    values = await communicator.txn_compare(compares, _txn)
    print(values)  # ['asdf']

compares = [
    CompareKey('cmpkey1').value == 'foo',
    CompareKey('cmpkey2').value < 'bar',
]
async with etcd.connect() with communicator:
    await communicator.put('failurekey', 'asdf')

    def _txn(success, failure):
        failure.get('failurekey')

    values = await communicator.txn_compare(compares, _txn)
    print(values)  # ['asdf']

If you don't need compare conditions for transaction, you can use EtcdCommunicator.txn(txn_builder), which is a shorthand for EtcdCommunicator.txn_compare([], lambda success, failure: txn_builder(success)).

async with etcd.connect() with communicator:
    def _txn(action):
        action.get('cmpkey1')
        action.get('cmpkey2')

    values = await communicator.txn(_txn)
    print(values)  # ['foo', 'baz']

Compiling Protobuf

$ scripts/compile_protobuf.py <target Etcd version>