wcs/list2form : filtrage des lignes vides

[progfou.git] / wcs / wcs-dynexport

#!/usr/bin/python
# -*- coding: utf-8 -*-
"""
Outil d'export dynamique 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 : 13 mars 2013

Depends: wcs, python-simplejson, python-magic

URL d'accès :
- /dynexport => liste des formulaires pour le domaine courant
- /dynexport/domains.json => liste des domaines disponibles
- /dynexport/formulaire => liste des options ci-dessous
- /dynexport/formulaire/fields.json
- /dynexport/formulaire/field-names.json
- /dynexport/formulaire/field-names.txt
- /dynexport/formulaire/data.json
- /dynexport/formulaire/last-run.log
- /dynexport/formulaire/liste-dossiers.json
- /dynexport/formulaire/clear-cache => vide le cache
- /dynexport/formulaire/data/nom-dossier.json
- /dynexport/formulaire/data/nom-dossier_attachement-1.xxx
- /dynexport/formulaire/data/nom-dossier_attachement-2.xxx
- /dynexport/formulaire/data/nom-dossier_attachement-…
"""
import sys
import os
import os.path
import logging
import time # time, gmtime, strftime, strptime, struct_time
import simplejson as json
import magic
import mimetypes
import unicodedata
from cStringIO import StringIO
from gzip import GzipFile
from re import match

DELAIS = 5 # maximum 5 secondes en cache
TIME_FORMAT = '%a, %d %b %Y %H:%M:%S GMT' # format date pour HTTP

WCS_ROOT_DIR = '/var/lib/wcs'
WCS_DOMAIN_SUFFIX = '.auf.org'
WCS_FORM_PREFIX = 'form-'
WCS_CACHE_DIR = '/var/tmp'

#--------------------------------------------------------------------------
# fonctions de traitement
#--------------------------------------------------------------------------

def http_redirect(location, code='302'):
    headers = {}
    headers['Content-Type'] = 'text/plain; charset=utf-8'
    headers['Status'] = '302 Redirection'
    headers['Location'] = location
    data = """If you see this, it means the automatic redirection has failed.
Please go to ${location}"""
    # envoi de la réponse
    headers = ''.join(map(lambda x: "%s: %s\r\n" % (x, headers[x]), headers))
    f = open('/dev/stdout', 'wb')
    f.write(headers + "\r\n")
    if data:
        f.write(data)
    f.flush()
    # arrêt du traitement
    sys.exit(0)

def http_reply_and_exit(data, mime_type='text/html', charset='utf-8'):
    # références horaires
    current_time = time.time()
    mtime = time.gmtime(current_time)
    etime = time.gmtime(current_time + DELAIS)
    if os.environ.has_key('HTTP_IF_MODIFIED_SINCE'):
        try:
            itime = time.strptime(os.environ['HTTP_IF_MODIFIED_SINCE'], TIME_FORMAT)
        except ValueError:
            itime = None
    else:
        itime = None
    # préparation des en-têtes et données
    headers = {}
    headers['Content-Type'] = '%s; charset=%s' % (mime_type, charset)
    headers['Last-Modified'] = time.strftime(TIME_FORMAT, mtime)
    headers['Expires'] = time.strftime(TIME_FORMAT, etime)
    if os.environ['REQUEST_METHOD'] == 'GET' and (not itime or mtime > itime):
        # détermination de la version demandée (compressée ou non)
        if os.environ.get('HTTP_ACCEPT_ENCODING','').split(',').count('gzip') > 0:
            zdata = StringIO()
            GzipFile('', 'w', 9, zdata).write(data)
            data = zdata.getvalue()
            headers['Content-Encoding'] = 'gzip'
        headers['Vary'] = 'Content-Encoding'
        headers['Content-Length'] = len(data)
    else:
        data = None
    # envoi de la réponse
    headers = ''.join(map(lambda x: "%s: %s\r\n" % (x, headers[x]), headers))
    f = open('/dev/stdout', 'wb')
    f.write(headers + "\r\n")
    if data:
        f.write(data)
    f.flush()
    # arrêt du traitement
    sys.exit(0)


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

    if type(s) is not unicode:
        s = unicode(s, 'utf-8')
    s = unicodedata.normalize('NFKD', s).encode('ASCII', 'ignore')
    r = ''
    for c in s:
        if ('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
    return r.strip(replacement_char)

def _make_wcs_cache_name(domain, form, name):
    return 'wcs-%s-%s-%s' % (domain, form, name)

def set_wcs_cache(domain, form, name, data):
    os.umask(0022)
    cache_filename = _make_wcs_cache_name(domain, form, name)
    f = open(os.path.join(WCS_CACHE_DIR, cache_filename), 'wb')
    f.write(data)
    f.close()

def get_wcs_cache(domain, form, name):
    data = None
    cache_filename = _make_wcs_cache_name(domain, form, name)
    cache_filename = os.path.join(WCS_CACHE_DIR, cache_filename)
    if os.path.exists(cache_filename):
        f = open(cache_filename, 'rb')
        data = f.read()
        f.close()
    return data

def clear_wcs_cache(domain, form):
    cache_filename = _make_wcs_cache_name(domain, form, '')
    for f in os.listdir(WCS_CACHE_DIR):
        if f.startswith(cache_filename):
            os.unlink(os.path.join(WCS_CACHE_DIR, f))

def get_wcs_domains():
    root = WCS_ROOT_DIR
    suffix = WCS_DOMAIN_SUFFIX
    try:
        l = os.listdir(root)
    except OSError:
        return None
    return [x for x in l if os.path.isdir(os.path.join(root, x)) and x.endswith(suffix)]

def get_wcs_forms(domain):
    root = os.path.join(WCS_ROOT_DIR, domain)
    prefix = WCS_FORM_PREFIX
    try:
        l = os.listdir(root)
    except OSError:
        return None
    return [x[len(prefix):] for x in l if os.path.isdir(os.path.join(root, x)) and x.startswith(prefix)]


def get_wcs_form_data(domain, form):
    """extraction des données du formulaire"""
    data = get_wcs_cache(domain, form, 'metadata.json')
    if data is not None:
        return json.loads(data, encoding='utf-8')
    # dictionnaire des metadonnées (qui seront mises en cache)
    metadata = {}

    os.umask(0022)
    logname = _make_wcs_cache_name(domain, form, 'last-run.log')
    logging.basicConfig(level=logging.DEBUG,
        format='%(asctime)s %(levelname)s %(message)s',
        filename=os.path.join(WCS_CACHE_DIR, logname),
        filemode='w')

    logging.info('Début.')

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

    pub = publisher.WcsPublisher.create_publisher()
    pub.app_dir = os.path.join(pub.app_dir, domain)
    formdef = FormDef.get_by_urlname(form)

    # nommage des champs de façon unique
    fields = {}
    field_names = {}
    field_names_duplicates = {}
    for field in formdef.fields:
        if isinstance(field, TitleField) or isinstance(field, CommentField):
            continue
        if field.varname:
            name = field.varname
        else:
            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})
        fields.update({field.id: {'name': field_names[field.id], 'label': field.label, 'varname': field.varname and field.varname or ''}})

    data = json.dumps(fields, ensure_ascii=False).encode('utf-8')
    set_wcs_cache(domain, form, 'fields.json', data)
    metadata.update({'fields': fields})

    # 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)
    magicmime.load()

    liste_dossiers = []
    liste_attachements = {}
    for object in formdef.data_class().select():
        if object.user is None:
            logging.warning("Dossier '%s' sans utilisateur associé ?!?"\
                            " On ignore...", object.id)
            continue

        result = {
            'num_dossier': object.id,
            'wcs_status': object.status,
            'wcs_workflow_status': (object.status.startswith('wf-') and \
                                object.get_workflow_status().name or None),
            'wcs_user_email': object.user.email,
            'wcs_user_display_name': object.user.display_name,
           #'wcs_last_modified': time.strftime('%Y-%m-%d %H:%M:%S', time.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
                    e_time = time.strftime('%Y-%m-%d %H:%M:%S', e.time)
                    comment = '%s -- %s %s' % (e.comment, who, e_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 = fields[field_id]['name']
            data = object.data.get(field_id)
            if data is None:
                result[field_name] = None
                continue
            if isinstance(field, StringField) or isinstance(field, TextField) \
            or isinstance(field, EmailField) or isinstance(field, ItemField):
                result[field_name] = data
            elif isinstance(field, ItemsField) or isinstance(field, TableField):
                result[field_name] = data # liste => peux-être joindre sur ';'
            elif isinstance(field, BoolField):
                result[field_name] = (data == 'True')
            elif isinstance(field, DateField):
                if isinstance(data, time.struct_time):
                    result[field_name] = '%04d-%02d-%02d' % (data.tm_year,
                                                    data.tm_mon, data.tm_mday)
                else:
                    result[field_name] = 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)
                    try:
                        #m = magicmime.from_file(p) => ce sera pour plus tard…
                        m = magicmime.file(p).split()[0].strip(';')
                        extension = mimetypes.guess_extension(m)
                    except:
                        logging.warning("Type de fichier inconnu pour '%s'.", p)
                        extension = None
                    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.warning("Type de champ inconnu '%s' pour '%s' (%s).",
                            field.__class__.__name__, field_name, field.label)

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

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

        # sauvegarde des chemins d'accès aux fichiers joints
        for f in qfiles:
            dst = filename + '_' + result[f]
            src = os.path.join(pub.app_dir, 'uploads', qfiles[f])
            liste_attachements.update({dst: src})

        # génération du fichier JSON
        data = json.dumps(result, ensure_ascii=False).encode('utf-8')
        set_wcs_cache(domain, form, 'data_%s.json' % filename, data)

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

    data = json.dumps(liste_attachements, ensure_ascii=False).encode('utf-8')
    set_wcs_cache(domain, form, 'data-files.json', data)
    metadata.update({'attachements': liste_attachements})

    liste_dossiers.sort()
    data = json.dumps(liste_dossiers, ensure_ascii=False).encode('utf-8')
    set_wcs_cache(domain, form, 'liste-dossiers.json', data)
    metadata.update({'dossiers': liste_dossiers})

    logging.info('Fin.')

    data = json.dumps(metadata, ensure_ascii=False).encode('utf-8')
    set_wcs_cache(domain, form, 'metadata.json', data)

#if __name__ == '__main__':
#    try:
#        extract_data(formdef, OUTPUT_DIRECTORY)
#    except:
#        logging.exception("Interruption du traitement pour cause d'erreur !")

#--------------------------------------------------------------------------
# gestion des requêtes web
#--------------------------------------------------------------------------

#l = []
#for k in sorted(os.environ):
#    l.append('%s=%s\n' % (k, os.environ[k]))
#data = ''.join(l)
#http_reply_and_exit(data, 'text/plain')

domain = os.environ.get('HTTP_HOST', '')
if domain not in get_wcs_domains():
    http_reply_and_exit("Domaine '%s' inconnu." % domain, 'text/plain')

path_info = os.environ.get('PATH_INFO', '')

path_prefix = os.environ.get('REQUEST_URI', '')
if len(path_info) > 0:
    path_prefix = path_prefix[:-len(path_info)]

if path_info == '':
    http_redirect(path_prefix + '/')

if path_info == '/':
    # liste des formulaires disponibles
    l = sorted(get_wcs_forms(domain))
    l = ['<li><a href="%s/">%s</a></li>' % (f, f) for f in l]
    title = '<p>Liste des formulaires disponibles&nbsp;:</p>\n'
    data = '<html>\n' + title + '<ul>\n' + '\n'.join(l) + '\n</ul>\n</html>'
    http_reply_and_exit(data, 'text/html')

if path_info == '/index.json':
    # liste des formulaires disponibles
    l = sorted(get_wcs_forms(domain))
    data = json.dumps(l, ensure_ascii=False, indent=' ').encode('utf-8')
    http_reply_and_exit(data, 'application/json')

if path_info == '/domains.json':
    # liste des domaines disponibles
    l = get_wcs_domains()
    data = json.dumps(l, ensure_ascii=False, indent=' ').encode('utf-8')
    http_reply_and_exit(data, 'application/json')

if match(r'^/[a-z0-9-]+$', path_info):
    http_redirect(path_prefix + path_info + '/')

if match(r'^/[a-z0-9-]+/$', path_info):
    form = path_info.split('/')[1]
    if form not in get_wcs_forms(domain):
        http_reply_and_exit("Formulaire '%s' inconnu." % form, 'text/plain')
    l = [ 'fields.json', 'field-names.json', 'field-names.txt', 'last-run.log', 'liste-dossiers.json' ]
    l = ['<li><a href="%s">%s</a></li>' % (f, f) for f in l]
    title = '<p>Liste des informations disponibles&nbsp;:</p>\n'
    action1 = """<p><a href="data/">Export des données</a></p>\n"""
    action2 = """<p><a href="clear-cache">Suppression du cache</a> (pour ré-export)</p>\n"""
    data = '<html>\n' + title + '<ul>\n' + '\n'.join(l) + '\n</ul>\n' + action1 + action2 + '</html>'
    http_reply_and_exit(data, 'text/html')

if match(r'^/[a-z0-9-]+/index.json$', path_info):
    form = path_info.split('/')[1]
    if form not in get_wcs_forms(domain):
        http_reply_and_exit("Formulaire '%s' inconnu." % form, 'text/plain')
    l = [ 'fields.json', 'field-names.json', 'field-names.txt', 'last-run.log', 'liste-dossiers.json', 'data', 'clear-cache' ]
    data = json.dumps(l, ensure_ascii=False, indent=' ').encode('utf-8')
    http_reply_and_exit(data, 'application/json')

if match(r'^/[a-z0-9-]+/clear-cache$', path_info):
    form = path_info.split('/')[1]
    if form not in get_wcs_forms(domain):
        http_reply_and_exit("Formulaire '%s' inconnu." % form, 'text/plain')
    clear_wcs_cache(domain, form)
    http_reply_and_exit('Ok.', 'text/plain')

if match(r'^/[a-z0-9-]+/fields.json$', path_info):
    form = path_info.split('/')[1]
    if form not in get_wcs_forms(domain):
        http_reply_and_exit("Formulaire '%s' inconnu." % form, 'text/plain')
    get_wcs_form_data(domain, form)
    d = json.loads(get_wcs_cache(domain, form, 'fields.json'), encoding='utf-8')
    data = json.dumps(d, ensure_ascii=False, indent=' ').encode('utf-8')
    http_reply_and_exit(data, 'application/json')

if match(r'^/[a-z0-9-]+/field-names.json$', path_info):
    form = path_info.split('/')[1]
    if form not in get_wcs_forms(domain):
        http_reply_and_exit("Formulaire '%s' inconnu." % form, 'text/plain')
    get_wcs_form_data(domain, form)
    d = json.loads(get_wcs_cache(domain, form, 'fields.json'), encoding='utf-8')
    d = dict([(k, d[k]['name']) for k in d])
    data = json.dumps(d, ensure_ascii=False, indent=' ').encode('utf-8')
    http_reply_and_exit(data, 'application/json')

if match(r'^/[a-z0-9-]+/field-names.txt$', path_info):
    form = path_info.split('/')[1]
    if form not in get_wcs_forms(domain):
        http_reply_and_exit("Formulaire '%s' inconnu." % form, 'text/plain')
    get_wcs_form_data(domain, form)
    d = json.loads(get_wcs_cache(domain, form, 'fields.json'), encoding='utf-8')
    d = [(k, d[k]['name'], d[k]['label']) for k in d]
    d = sorted(d, key=lambda x: int(x[0]))
    text = u''.join([u'%s:%s:%s\n' % (x[0], x[1], x[2]) for x in d])
    data = text.encode('utf-8')
    http_reply_and_exit(data, 'text/plain')

if match(r'^/[a-z0-9-]+/last-run.log$', path_info):
    form = path_info.split('/')[1]
    if form not in get_wcs_forms(domain):
        http_reply_and_exit("Formulaire '%s' inconnu." % form, 'text/plain')
    get_wcs_form_data(domain, form)
    data = get_wcs_cache(domain, form, 'last-run.log')
    http_reply_and_exit(data, 'text/plain')

if match(r'^/[a-z0-9-]+/liste-dossiers.json$', path_info):
    form = path_info.split('/')[1]
    if form not in get_wcs_forms(domain):
        http_reply_and_exit("Formulaire '%s' inconnu." % form, 'text/plain')
    get_wcs_form_data(domain, form)
    data = json.loads(get_wcs_cache(domain, form, 'liste-dossiers.json'), encoding='utf-8')
    data = json.dumps(data, ensure_ascii=False, indent=' ').encode('utf-8')
    http_reply_and_exit(data, 'application/json')

if match(r'^/[a-z0-9-]+/data$', path_info):
    http_redirect(path_prefix + path_info + '/')

if match(r'^/[a-z0-9-]+/data/$', path_info):
    form = path_info.split('/')[1]
    if form not in get_wcs_forms(domain):
        http_reply_and_exit("Formulaire '%s' inconnu." % form, 'text/plain')
    get_wcs_form_data(domain, form)
    dossiers = json.loads(get_wcs_cache(domain, form, 'liste-dossiers.json'), encoding='utf-8')
    attachements = json.loads(get_wcs_cache(domain, form, 'data-files.json'), encoding='utf-8')
    l = sorted(dossiers + attachements.keys())
    if len(l) > 0:
        l = ['<li><a href="%s">%s</a></li>' % (f, f) for f in l]
        title = '<p>Liste des documents disponibles&nbsp;:</p>\n'
        data = '<html>\n' + title + '<ul>\n' + '\n'.join(l) + '\n</ul>\n</html>'
    else:
        data = '<html>\n<p>Aucun document disponible.</p>\n</html>'
    http_reply_and_exit(data, 'text/html')

if match(r'^/[a-z0-9-]+/data/index.json$', path_info):
    form = path_info.split('/')[1]
    if form not in get_wcs_forms(domain):
        http_reply_and_exit("Formulaire '%s' inconnu." % form, 'text/plain')
    get_wcs_form_data(domain, form)
    dossiers = json.loads(get_wcs_cache(domain, form, 'liste-dossiers.json'), encoding='utf-8')
    attachements = json.loads(get_wcs_cache(domain, form, 'data-files.json'), encoding='utf-8')
    l = sorted(dossiers + attachements.keys())
    data = json.dumps(l, ensure_ascii=False, indent=' ').encode('utf-8')
    http_reply_and_exit(data, 'application/json')

if match(r'^/[a-z0-9-]+/data/[^/]+$', path_info):
    form = path_info.split('/')[1]
    if form not in get_wcs_forms(domain):
        http_reply_and_exit("Formulaire '%s' inconnu." % form, 'text/plain')
    get_wcs_form_data(domain, form)
    doc = path_info.split('/')[3]
    dossiers = json.loads(get_wcs_cache(domain, form, 'liste-dossiers.json'), encoding='utf-8')
    if doc in dossiers:
        data = get_wcs_cache(domain, form, 'data_' + doc)
        data = json.loads(data, encoding='utf-8')
        data = json.dumps(data, ensure_ascii=False, indent=' ').encode('utf-8')
        http_reply_and_exit(data, 'application/json')
    attachements = json.loads(get_wcs_cache(domain, form, 'data-files.json'), encoding='utf-8')
    if doc in attachements:
        data = open(attachements[doc], 'rb').read()
        mime_type = mimetypes.guess_type(doc)[0]
        if mime_type is None:
            mime_type = 'application/octet-stream'
        http_reply_and_exit(data, mime_type)
    http_reply_and_exit("Document '%s' inconnu." % path_info, 'text/plain')

http_reply_and_exit("Requête '%s' inconnue." % path_info, 'text/plain')