#!/usr/bin/env python
'''timbl.py - Using Timbl through Python

This module contains a function timbl() which is a simple
interface to TiMBL.

Copyright (c) 2011 CLiPS. All rights reserved.

# License: GNU General Public License, see http://www.clips.ua.ac.be/~vincent/scripts/LICENSE.txt
'''
__version__="1.1"
__author__="Vincent Van Asch"
__date__="August 2011"

import sys
import os
import subprocess
import time

# CONSTANTS #########################################################################

_timbl = '~/bin/Timbl'
NO = 0
TIME = 1
SETUP = 2
OUTPUT = 3
ALL = 4

# BODY ##############################################################################

class TiMBLError(Exception): pass


def abs(p, check=False):
    if not isinstance(p, (str, unicode)): return p
    if p:
        path = os.path.abspath(os.path.expanduser(p))
        if check and not os.path.isfile(path): raise TiMBLError('%s does not exist' %path)
        return path
        

def log(s): print >>sys.stderr, '%s: %s' %(time.strftime('%d/%m/%Y %H:%M:%S'), s)

def humantime(secs):
    '''Returns seconds as (days, hours, minutes, seconds)'''
    days = int(secs/(24*3600))
    hours=int(secs%(24*3600)/3600)
    minutes=int((secs%3600)/60)
    seconds=int((secs%3600)%60)
    return [days, hours, minutes, seconds]


def timbl(trainingfile, testfile, options=[], outputfile=None, stderr=None, stdout=None, binary=_timbl, verbose=NO):
    '''
    trainingfile: the path to a file with labeled instances to use as training data
    testfile: the path to a file with (un)labeled instances to use as test data
    options: a list of Timbl options. E.g. ['-k10', '-dIL', '-v', 'cs+cm']
    outputfile: the path to a file to write the labeled test instances to. If None, output is written to the
                default Timbl output file.
    
    stderr: the path to a file to write stderr output to or an open file object. If None, the output is not saved.
    stderr: the path to a file to write stdout output to or an open file object. If None, the output is not saved.

    binary: the path to a working Timbl binary (default: ~/bin/Timbl)
    verbose: to set the amount of information printed to stdout. 
             Levels:
              NO : nothing (default)
              TIME : time taken to run TiMBL
              SETUP : all setup data is also printed
              OUTPUT : time to run TiMBL and the stdout of the process
              ALL: everything
    '''
    #Absolute paths
    trainingfile = abs(trainingfile, check=True)
    testfile = abs(testfile, check=True)
    outputfile = abs(outputfile)
    stderr = abs(stderr)
    stdout = abs(stdout)
    binary = abs(binary)
    
    # Check the binary
    if not os.access(binary, os.X_OK): raise TiMBLError('%s is not a working TiMBL binary' %(binary))
    
    # Create the command
    cmd = [binary, '-f', trainingfile, '-t', testfile]
    if outputfile: cmd.extend(['-o', outputfile])
    cmd.extend(options)

    # Set up
    if verbose > 1:
        print '''#########################################################
TRAIN  : %s
TEST   : %s
OPTIONS: %s
OUTPUT : %s

STDOUT : %s
STDERR : %s

TIMBL  : %s
FULL COMMAND:\n %s
#########################################################''' %(trainingfile, testfile, ' '.join(options) or 'default', outputfile or 'default', stdout or '-', stderr or '-', binary, ' '.join(cmd))

    
    # Keep starttime
    start = time.time()
    
    proc = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
    out, err = proc.communicate()

    if verbose > 2: print 'STDOUT\n', out, '\n#########################################################'
    if verbose > 3: print 'STDERR\n', err, '\n#########################################################'
    if verbose > 0 and not proc.returncode:
        elapsed = humantime(time.time() - start)
        string='Timbl fully completed in'
        steps = ['day', 'h', 'min', 's']
        for name, value in zip(steps, elapsed):
            if value == 0 and string.endswith('n'): continue 
            if name == 'day' and value > 1: name+='s'
            string += (' %d%s' %(value, name))
        if string.endswith('n'): string+=' less than a second.'
        log(string)
        
    if stdout:
        if isinstance(stdout, file):
            if stdout.closed:
                raise TiMBLError('stdout file object is closed')
            else:
                stdout.write(out)
        else:
            with open(stdout, 'w') as f:
                f.write(out)
    if stderr:
        if isinstance(stderr, file):
            if stderr.closed:
                raise TiMBLError('stderr file object is closed')
            else:
                stderr.write(err)
        else:
            with open(stderr, 'w') as f:
                f.write(err)
    
    if proc.returncode:
        print >>sys.stderr, err
        print >>sys.stderr, out
        raise TiMBLError('TiMBL did not terminate correctly (return code: %s)' %str(proc.returncode))