whs / cjlib

Codejam helper library

awkwin's Codejam Library

This is my library collection used in coding competitions.

License

All codes except for mathapi is licensed under StealItPl 1.1.

API

• cjlib.baseN(num, base[, numerals]): Convert num to base base with numerals as character representatives of values (default to 0-9a-z) source
• cjlib.mathapi: @neizod's mathapi. Source and docs

input.py

Input library. This library only works with only one input at a time. (i.e. input.get should be called only once)

• input.get([default[, skip=True]]): Read input from first command line argument or use default. Usually the first line is the lines count and will be automatically chopped. Disable this by setting skip=False.
• input.line(): Return next line from the input. (No line ending is included)
• input.lines([cnt=2]): Return next cnt lines from the input as an array.
• input.block(): Return next block in the same format as input.lines without the line number count. A block is defined as a line with a number showing how many lines are in this block. Example of a block:

4
2 2 2
2 1 2
2 1 2
2 2 2

Returns ["2 2 2", "2 1 2", "2 1 2", "2 2 2"].

• input.neof(): Return True if there is still more input. Usually used in a while loop.

runner.py

Task multiprocessing.

Usage:

from cjlib.input import neof, line, get
from cjlib.runner import TaskRunner, DummyRunner

def is_odd(inp):
	return inp%2 == 1

r = TaskRunner(is_odd, DummyRunner)

get("""4
1
3
4
5
""")

while neof():
	r.add(int(line()))

r.run()

• The second argument to TaskRunner is optional and use to specify multiprocessing backend. Defaults to ThreadRunner. Available modules are:
  • DummyRunner: Does not use multiprocessing. Easiest to debug.
  • MPQRunner: Use multiprocessing.Pool. Fastest but requires more memory.
  • ThreadRunner: Use threading.Thread. (Actually an MPQRunner with multiprocessing.dummy.Pool as thread pool)
• MPQRunner use 4 as hard-coded subprocess count. Supply processCnt=8 to TaskRunner when using MPQRunner to use 8 processes and so on.
• The output will be returned in stdout after all backend has finished processing. Status messages are sent to Python's logging module with info level and usually be sent to stderr.
• To add case number to output, use r.run(True). The output will be in Codejam format: Case #n: out
• Accessing TaskRunner in the processing function (in this case, is_odd) is not safe.