Assure le chemin de wcs-extract.

[progfou.git] / wcs / wcs-extract

#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
Outil d'export de données w.c.s.

Copyright : Agence universitaire de la Francophonie — www.auf.org
Licence : GNU General Public Licence, version 2
Auteur : Jean Christophe André
Date de création : 15 octobre 2009

Depends: wcs, python-simplejson, python-magic
"""
import os
import os.path
import shutil
import logging
from time import gmtime, strftime
import simplejson as json
import magic
import mimetypes

from wcs import publisher
from wcs.formdef import FormDef
from wcs.fields import TitleField, CommentField, TextField, \
                       StringField, ItemField, EmailField, \
                       DateField, FileField, BoolField


def reduce_to_alnum(s, replacement_char='-'):
    """réduction d'une chaîne de caractères à de l'alpha-numérique"""

    avec_accent = u'çÇáàâÁÀÂéèêëÉÈÊËíìîïÍÌÎÏóòôöÓÒÔÖúùûüÚÙÛÜýỳyÿÝỲYŸ'
    sans_accent = u'cCaaaAAAeeeeEEEEiiiiIIIIooooOOOOuuuuUUUUyyyyYYYY'
    if type(s) is not unicode:
        s = unicode(s, 'utf-8')
        u  = False
    r = ''
    for c in s:
        index = avec_accent.find(c)
        if index >= 0:
            r += sans_accent[index]
        elif ('a' <= c.lower() <= 'z') or ('0' <= c <= '9'):
            r += c
        elif len(r) > 0 and r[-1] != replacement_char:
            r += replacement_char
        else: # r == '' or r[-1] == replacement_char
            pass
    r = r.strip(replacement_char)
    if not u:
        r = r.encode('utf-8')
    return r


def extract_fields(formdef, output_directory):
    """nommage des champs de façon unique"""
    # TODO: devrait retourner un résultat, qui serait alors sauvé en dehors

    # XXX: hack temporaire… :-/
    global field_names

    f = open(os.path.join(output_directory, 'field-names.txt'), 'w')

    field_names = {}
    field_names_duplicates = {}
    for field in formdef.fields:
        if isinstance(field, TitleField) or isinstance(field, CommentField):
            continue
        name = reduce_to_alnum(field.label,'_').lower()
        if name in field_names.values(): # duplicat
            field_names_duplicates[name] = field_names_duplicates.get(name, 1) + 1
            name = '%s_%d' % (name, field_names_duplicates[name])
        field_names.update({field.id: name})
        print >>f, "%s:%s:%s" % (field.id, field_names[field.id], field.label)

    f.close()

    f = open(os.path.join(output_directory, 'field-names.json'), 'wb')
    f.write(json.dumps(field_names, ensure_ascii=False, sort_keys=True))
    f.close()


def extract_data(formdef, output_directory):
    """extraction des données du formulaire"""
    # TODO: devrait retourner un résultat, qui serait alors sauvé en dehors

    # XXX: hack temporaire… :-/
    global pub

    # on charge la base des types MIME une fois pour toutes
    #magicmime = magic.Magic(mime=True) => ce sera pour plus tard…
    magicmime = magic.open(magic.MAGIC_MIME_TYPE)

    liste_dossiers = []
    for object in formdef.data_class().select():
        result = {
            'num_dossier': object.id,
            'wcs_status': object.status,
            'wcs_workflow_status': object.get_workflow_status().name,
            'wcs_user_email': object.user.email,
            'wcs_user_display_name': object.user.display_name,
           #'wcs_last_modified': strftime('%Y-%m-%d %H:%M:%S', gmtime(object.last_modified())),
            'wcs_comments': [],
        }

        if object.evolution is not None:
            for e in object.evolution:
                if e.comment is not None:
                    who = pub.user_class.get(e.who).display_name
                    time = strftime('%Y-%m-%d %H:%M:%S', e.time)
                    comment = '%s -- %s %s' % (e.comment, who, time)
                    result['wcs_comments'].append(comment)

        qfiles = { }
        for field in formdef.fields:
            field_id = str(field.id)
            if not field_id in object.data:
                continue
            if isinstance(field, TitleField) or isinstance(field, CommentField):
                continue
            field_name = field_names[field_id]
            data = object.data.get(field_id)
            if isinstance(field, StringField) or isinstance(field, TextField) \
            or isinstance(field, EmailField) or isinstance(field, ItemField):
                result[field_name] = data
            elif isinstance(field, BoolField):
                result[field_name] = (data == 'True')
            elif isinstance(field, DateField):
                result[field_name] = strftime('%Y-%m-%d', data)
            elif isinstance(field, FileField):
                if '.' in data.orig_filename:
                    extension = data.orig_filename.rpartition('.')[2].lower()
                else: # il n'y a pas d'extension dans le nom de fichier
                    p = os.path.join(pub.app_dir, 'uploads', data.qfilename)
                    #m = magicmime.from_file(p) => ce sera pour plus tard…
                    m = magicmime.file(p)
                    extension = mimetypes.guess_extension(m)
                    if extension is not None:
                        extension = extension[1:]
                    else:
                        extension = 'unknown'
                result[field_name] = "%s.%s" % (field_name, extension)
                qfiles[field_name] = data.qfilename
            else:
                logging.error("Type de champ inconnu '%s' pour '%s'.",
                                    field.__class__.__name__, field.label)
                raise RuntimeError

        num_dossier = result['num_dossier']
        nom = reduce_to_alnum(result['nom']).upper()
        prenom = reduce_to_alnum(result['prenom']).upper()
        adel = result['adresse_electronique'].replace('@','-').lower()

        filename = "%04d-%s-%s-%s" % (num_dossier, nom, prenom, adel)
        liste_dossiers.append(filename + '.json')

        # copie des fichiers joints
        for f in qfiles:
            result[f] = filename + '_' + result[f]
            src = os.path.join(pub.app_dir, 'uploads', qfiles[f])
            dst = os.path.join(output_directory, 'data', result[f])
            if not os.path.exists(dst) or os.path.getmtime(src) > os.path.getmtime(dst):
                shutil.copy2(src, dst)
                os.chmod(dst, 0644)

        # génération du fichier JSON
        jsonname = os.path.join(output_directory, 'data', filename + '.json')
        f = open(jsonname, 'wb')
        f.write(json.dumps(result, ensure_ascii=False, sort_keys=True))
        f.close()

        logging.info("Dossier '%s' : %s.",
                                    filename, result['wcs_workflow_status'])

    f = open(os.path.join(output_directory, 'liste-dossiers.json'), 'wb')
    f.write(json.dumps(liste_dossiers, ensure_ascii=False, sort_keys=True))
    f.close()


if __name__ == '__main__':
    import sys

    if len(sys.argv) != 4:
        print >>sys.stderr, "Usage : %s <dossier-destination> <site> <formulaire>" % sys.argv[0]
        sys.exit(1)

    VHOST = sys.argv[2]
    FORM_NAME = sys.argv[3]
    OUTPUT_DIRECTORY = os.path.join(sys.argv[1], VHOST, FORM_NAME)

    os.umask(0022)
    # création du dossier d'extraction, au besoin
    if not os.path.isdir(OUTPUT_DIRECTORY):
        os.makedirs(OUTPUT_DIRECTORY, 0755)

    logging.basicConfig(level=logging.DEBUG,
        format='%(asctime)s %(levelname)s %(message)s',
        filename=os.path.join(OUTPUT_DIRECTORY, 'last-run.log'),
        filemode='w')

    logging.info('Début.')

    pub = publisher.WcsPublisher.create_publisher()
    pub.app_dir = os.path.join(pub.app_dir, VHOST)

    formdef = FormDef.get_by_urlname(FORM_NAME)

    extract_fields(formdef, OUTPUT_DIRECTORY)

    extract_data(formdef, OUTPUT_DIRECTORY)

    logging.info('Fin.')