Skip to main content

Utilities for lazy iterables

Project description


clj is a collection of functions to work with lazy iterators in Python.

It’s a module for those times when you did too much Clojure and came back to Python wondering where are all these distinct, drop-while, cycle, first, etc.

The library is oriented toward laziness and performance. Functions are implemented with as little overhead as possible.


pip install clj

Since 0.2.0, only Python 3.7+ is supported.


The library is considered stable and is used in production.


;; Clojure
(->> coll (map inc) (filter even?) distinct count println)
; which expands to:
(println (count (distinct (filter even? (map inc coll)))))
# Python
from clj import count, distinct, inc, is_even

print(count(distinct(filter(is_even, map(inc, coll)))))

Note that count() works on both sequences in generators; in the latter case it doesn’t load everything in memory like e.g. len(list(g)) would do.

Core Ideas

  • Lazy by default. All functions should work on arbitrary iterators and return generators.
  • This is Python. We keep Python’s semantics instead of trying to reproduce Clojure in Python (e.g. 0 and [] are logically true in Clojure but false in Python; None is not equivalent to an empty collection).
  • Don’t Reinvent the Wheel. We don’t reimplement built-in functions unless they miss something, like range that can’t be called without argument to yield an infinite sequence.


The general naming scheme is: use underscores instead of hyphens; start the function with is_ if its Clojure counterparts ends with a ?.


We aim to implement all Clojure functions that operate on sequences (see the list here). They all work on iterables and return generators by default (Python’s closest equivalent of lazy seqs). We don’t support transducers.

Clojure clj Comment
distinct distinct
filter filter Alias to Python’s built-in filter.
remove remove
keep keep
keep-indexed keep_indexed
cons cons
concat concat Equivalent to itertools.chain.
lazy-cat - Use Python’s itertools.chain.
mapcat mapcat
cycle cycle
interleave interleave
interpose interpose
rest rest
next - Use rest.
fnext - Use second.
nnext - Use rest(rest(…))
drop drop
drop-while drop_while Equivalent to itertools.dropwhile.
nthnext - Use drop.
take take
take-nth take_nth
take-while take_while Equivalent to itertools.takewhile.
butlast butlast
drop-last drop_last
flatten flatten
reverse reverse
sort - Use Python’s built-in sort.
sort-by - Use sort(…, key=your_function).
shuffle shuffle
split-at split_at
split-with split_with
partition partition (partition n step pad coll) becomes partition(coll, n, step, pad). Only the case step=n is supported for now.
partition-by partition_by
map map Alias to Python’s built-in map.
pmap -
replace replace
reductions reductions (reductions f i c) becomes reductions(f, c, i).
map-indexed map_indexed
seque -
first first None is not a valid parameter.
ffirst ffirst None is not a valid parameter.
nfirst nfirst
second second
nth nth
last last
rand-nth - Use Python’s random.choice.
zipmap zipmap
into -
reduce - Use Python’s functools.reduce.
set - Use Python’s set.
vec - Use Python’s list.
into-array - Use Python’s list.
to-array-2d -
frequencies - Use Python’s collections.Counter.
group-by group_by
apply - Use the f(*args) construct.
not-empty -
some some
seq? is_seq
every? every
not-every? not_every
not-any? not_any
empty? -
empty empty
doseq - Use for … in.
dorun dorun
doall - Use Python’s list.
realized? -
seq - Use Python’s list.
vals - Use Python’s dict.values.
keys - Use Python’s dict.keys.
rseq -
repeatedly repeatedly
iterate iterate
repeat repeat (repeat n x) becomes repeat(x, n). Equivalent to itertools.repeat.
range range Prefer Python’s range for everything but infinite generators.
line-seq - Loop over an io.BufferedReader.
resultset-seq -
re-seq - Use Python’s re.finditer.
tree-seq tree_seq
file-seq - Use Python’s os.walk.
xml-seq -
iterator-seq -
enumeration-seq -
hash-map - Use Python’s dict.
array-map - Use Python’s dict.
sorted-map - Use collections.OrderedDict.
hash-set - Use Python’s set.
set - Use Python’s set.
dedupe dedupe

We also implemented count, which uses Python’s len when possible and fallbacks on a for loop for other cases.


We also provide miscellaneous functions as well as functions that work on functions.

Clojure clj Comment
identity identity
partial - Use Python’s functools.partial
comp comp
complement complement
constantly constantly
juxt juxt
distinct? is_distinct
Clojure clj Comment
inc inc
dec dec

Project details

Download files

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

Source Distribution

clj-0.3.1.tar.gz (12.9 kB view hashes)

Uploaded source

Built Distribution

clj-0.3.1-py3-none-any.whl (11.9 kB view hashes)

Uploaded 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