Skip to main content

Python I/O pipe utilities

Project description Code Health

Tubing is a Python I/O library. What makes tubing so freakin’ cool is the gross abuse of the bit-wise OR operator (|). Have you ever been writing python code and thought to yourself, “Man, this is great, but I really wish it was a little more like bash.” Whelp, we’ve made python a little more like bash.If you are a super lame nerd-kid, you can replace any of the bit-wise ORs with the tube() function and pray we don’t overload any other operators in future versions. Here’s how you install tubing:

$ pip install tubing

Tubing is pretty bare-bones at the moment. We’ve tried to make it easy to add your own functionality. Hopefully you find it not all that unpleasant. There are three sections below for adding sources, tubes and sink. If you do make some additions, think about committing them back upstream. We’d love to have a full suite of tools.

Now, witness the power of this fully operational I/O library.

from tubing import sources, tubes, sinks

objs = [
        name="Bob Corsaro",
        name="Tom Brady",
sources.Objects(objs) \
     | tubes.JSONDumps() \
     | tubes.Joined(by=b"\n") \
     | tubes.Gzip() \
     | sinks.File("output.gz", "wb")

Then in our old friend bash.

$ zcat output.gz
{"alignment": "evil", "birthdate": "08/03/1977", "name": "Bob Corsaro"}
{"alignment": "good", "birthdate": "08/03/1977", "name": "Tom Brady"}

You can find more documentation on readthedocs




Takes a list of python objects.


Creates a stream from a file.


Takes a byte string.


Takes an object with a read function.


Takes an addr, port and socket() args. .


Takes an method, url and any args that can be passed to requests library.



Unzips a binary stream.


Zips a binary stream.


Parses a byte string stream of raw JSON objects. Will try to use ujson, then built-in json.


Serializes an object stream using json.dumps. Will try to use ujson, then built-in json.


Splits a stream that supports the split method.


Joins a stream of the same type as the by argument.


Takes a sink and passes chunks along apparatus.


Takes a transformer function for single items in stream.


Takes a filter test callback and only forwards items that pass.


Takes a transformer function for batch of stream items.



A list that stores all passed items to self.


Saves each chunk self.results.


Writes each chunk to a file.


Writes data via HTTPPost.


Takes algorithm name, updates hash with contents.


Writes each chunk to the tubing.tubes debugger with level DEBUG.



Create stream from an S3 object.


Stream data to S3 object.


Stream elasticsearch.DocUpdate objects to the elasticsearch _bulk endpoint.


To make your own source, create a Reader class with the following interface.

class MyReader(object):
    MyReader returns count instances of data.
    def __init__(self, data="hello world\n", count=10): = data
        self.count = count

    def read(self, amt):
        read(amt) returns $amt of data and a boolean indicating EOF.
        if not amt:
            amt = self.count
        r = * min(amt, self.count)
        self.count -= amt
        return r, self.count <= 0

The important thing to remember is that your read function should return an iterable of units of data, not a single piece of data. Then wrap your reader in the loving embrace of MakeSourceFactory.

from tubing import sources

MySource = sources.MakeSourceFactory(MyReader)

Now it can be used in a apparatus!

from __future__ import print_function

from tubing import tubes
sink = MySource(data="goodbye cruel world!", count=1) \
     | tubes.Joined(by=b"\n") \
     | sinks.Bytes()

# Output: goodbye cruel world!


Making your own tube is a lot more fun, trust me. First make a Transformer.

class OptimusPrime(object):
    def transform(self, chunk):
        return list(reversed(chunk))

chunk is an iterable with a len() of whatever type of data the stream is working with. In Transformers, you don’t need to worry about buffer size or closing or exception, just transform an iterable to another iterable. There are lots of examples in

Next give Optimus Prime a hug.

from tubing import tubes

AllMixedUp = tubes.MakeTranformerTubeFactory(OptimusPrime)

Ready to mix up some data?

from __future__ import print_function

import json
from tubing import sources, sinks

objs = [{"number": i} for i in range(0, 10)]

sink = sources.Objects(objs) \
     | AllMixedUp(chunk_size=2) \
     | sinks.Objects()

# Output: [{"number": 1}, {"number": 0}, {"number": 3}, {"number": 2}, {"number": 5}, {"number": 4}, {"number": 7}, {"number": 6}, {"number": 9}, {"number": 8}]


Really getting tired of making documentation… Maybe I’ll finish later. I have real work to do.

Well.. I’m this far, let’s just push through.

from __future__ import print_function
from tubing import sources, tubes, sinks

class StdoutWriter(object):
    def write(self, chunk):
        for part in chunk:

    def close(self):
        # this function is optional
        print("That's all folks!")

    def abort(self):
        # this is also optional
        print("Something terrible has occurred.")

Debugger = sinks.MakeSinkFactory(StdoutWriter)

objs = [{"number": i} for i in range(0, 10)]

sink = sources.Objects(objs) \
     | AllMixedUp(chunk_size=2) \
     | tubes.JSONDumps() \
     | tubes.Joined(by=b"\n") \
     | Debugger()
# Output:
#{"number": 1}
#{"number": 0}
#{"number": 3}
#{"number": 2}
#{"number": 5}
#{"number": 4}
#{"number": 7}
#{"number": 6}
#{"number": 9}
#{"number": 8}
#That's all folks!

Release history Release notifications | RSS feed

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release. See tutorial on generating distribution archives.

Built Distribution

tubing-0.0.2.post180-py2.py3-none-any.whl (23.5 kB view hashes)

Uploaded py2 py3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page