views.py 19.7 KB
Newer Older
1 2 3
"""
    Routes
"""
4
import copy
Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
5
import io
Pierre Dittgen's avatar
Pierre Dittgen committed
6
import itertools
7
import json
8
import logging
9
import subprocess
10
import tempfile
11
from datetime import datetime
12
from operator import itemgetter
13 14
from pathlib import Path
from urllib.parse import quote_plus
15

Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
16
import requests
17
from backports.datetime_fromisoformat import MonkeyPatch
18
from commonmark import commonmark
Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
19
from flask import make_response, redirect, render_template, request, url_for
Pierre Dittgen's avatar
Pierre Dittgen committed
20 21
from validata_core import compute_badge, csv_helpers, messages
from validata_core.loaders import custom_loaders
22

Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
23 24
import tabulator

25 26 27 28
from . import app, config
from .ui_util import flash_error, flash_warning
from .validata_util import ValidataSource
from .validate_helper import ValidatorHelper
29

30 31
MonkeyPatch.patch_fromisoformat()

32 33
log = logging.getLogger(__name__)

34 35

def extract_source_data(source: ValidataSource, preview_rows_nb=5):
36
    """ Computes table preview """
37 38 39

    def stringify(val):
        """ Transform value into string """
Pierre Dittgen's avatar
Pierre Dittgen committed
40
        return '' if val is None else str(val)
41

42 43
    header = None
    rows = []
Pierre Dittgen's avatar
Pierre Dittgen committed
44
    nb_rows = 0
45

Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
46 47 48 49 50 51
    # if source.scheme == 'file':
    #     source.scheme = 'bytes'

    import ipdb
    ipdb.set_trace()

52
    options = {}
53
    if source.format == "csv":
54
        options['delimiter'] = csv_helpers.detect_dialect(source.source, format=source.format, scheme=source.scheme,
55
                                                          custom_loaders=custom_loaders).delimiter
56
    with tabulator.Stream(source.source, format=source.format, scheme=source.scheme, custom_loaders=custom_loaders,
57
                          **options) as stream:
58 59
        for row in stream:
            if header is None:
60
                header = ['' if v is None else v for v in row]
61
            else:
62
                rows.append(list(map(stringify, row)))
Pierre Dittgen's avatar
Pierre Dittgen committed
63
                nb_rows += 1
64
    preview_rows_nb = min(preview_rows_nb, nb_rows)
65 66
    return {'header': header,
            'rows_nb': nb_rows,
67 68 69
            'data_rows': rows,
            'preview_rows_nb': preview_rows_nb,
            'preview_rows': rows[:preview_rows_nb]}
70 71


Pierre Dittgen's avatar
Pierre Dittgen committed
72 73
def improve_errors(errors):
    """Add context to errors, converts markdown content to HTML"""
74

Pierre Dittgen's avatar
Pierre Dittgen committed
75 76
    def improve_err(err):
        """Adds context info based on row-nb presence and converts content to HTML"""
77

Pierre Dittgen's avatar
Pierre Dittgen committed
78 79 80 81
        # Context
        update_keys = {
            'context': 'body' if 'row-number' in err and not err['row-number'] is None else 'table',
        }
Pierre Dittgen's avatar
Pierre Dittgen committed
82

Pierre Dittgen's avatar
Pierre Dittgen committed
83
        # markdown to HTML (with default values for 'title' and 'content')
Pierre Dittgen's avatar
Pierre Dittgen committed
84

Pierre Dittgen's avatar
Pierre Dittgen committed
85 86 87
        # Set default title if no title
        if not 'title' in err:
            update_keys['title'] = '[{}]'.format(err['code'])
Pierre Dittgen's avatar
Pierre Dittgen committed
88

Pierre Dittgen's avatar
Pierre Dittgen committed
89 90 91 92
        # Convert message to markdown only if no content
        # => for pre-checks errors
        if 'message' in err and not 'content' in err:
            update_keys['message'] = commonmark(err['message'])
Pierre Dittgen's avatar
Pierre Dittgen committed
93

Pierre Dittgen's avatar
Pierre Dittgen committed
94 95 96
        # Else, default message
        elif not 'message' in err or err['message'] is None:
            update_keys['message'] = '[{}]'.format(err['code'])
Pierre Dittgen's avatar
Pierre Dittgen committed
97

Pierre Dittgen's avatar
Pierre Dittgen committed
98 99 100
        # Message content
        md_content = '*content soon available*' if not 'content' in err else err['content']
        update_keys['content'] = commonmark(md_content)
101

Pierre Dittgen's avatar
Pierre Dittgen committed
102
        return {**err, **update_keys}
103

Pierre Dittgen's avatar
Pierre Dittgen committed
104
    return list(map(improve_err, errors))
105 106


Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
107
def create_validata_ui_report(validata_core_report, schema_dict):
108 109 110 111 112 113
    """ Creates an error report easier to handle and display in templates:
        - only one table
        - errors are contextualized
        - error-counts is ok
        - errors are grouped by lines
        - errors are separated into "structure" and "body"
114
        - error messages are improved
115
    """
Pierre Dittgen's avatar
Pierre Dittgen committed
116
    report = copy.deepcopy(validata_core_report)
117 118 119 120 121 122 123 124 125

    # One table is enough
    del report['table-count']
    report['table'] = report['tables'][0]
    del report['tables']
    del report['table']['error-count']
    del report['table']['time']
    del report['table']['valid']
    del report['valid']
126 127 128
    # use _ instead of - to ease information picking in jinja2 template
    report['table']['row_count'] = report['table']['row-count']

Pierre Dittgen's avatar
Pierre Dittgen committed
129 130 131 132
    # Handy col_count info
    headers = report['table'].get('headers', [])
    report['table']['col_count'] = len(headers)

Pierre Dittgen's avatar
Pierre Dittgen committed
133
    # Computes column info
Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
134 135
    fields_dict = {f['name']: (f.get('title', 'titre non défini'), f.get('description', ''))
                   for f in schema_dict.get('fields', [])}
Pierre Dittgen's avatar
Pierre Dittgen committed
136 137 138
    report['table']['headers_title'] = [fields_dict[h][0] if h in fields_dict else 'colonne inconnue' for h in headers]
    report['table']['headers_description'] = [fields_dict[h][1]
                                              if h in fields_dict else 'Cette colonne n\'est pas définie dans le schema' for h in headers]
139

140
    # Provide better (french) messages
Pierre Dittgen's avatar
Pierre Dittgen committed
141 142
    errors = improve_errors(report['table']['errors'])
    del report['table']['errors']
143

144 145 146 147 148 149 150
    # Count errors
    report['error_count'] = len(errors)
    del report['error-count']

    # Then group them in 2 groups : structure and body
    report['table']['errors'] = {'structure': [], 'body': []}
    for err in errors:
151
        if err['tag'] == 'structure':
152 153 154 155
            report['table']['errors']['structure'].append(err)
        else:
            report['table']['errors']['body'].append(err)

Pierre Dittgen's avatar
Pierre Dittgen committed
156
    # Checks if there are structure errors different to invalid-column-delimiter
Pierre Dittgen's avatar
Pierre Dittgen committed
157 158 159
    structure_errors = report['table']['errors']['structure']
    report['table']['do_display_body_errors'] = len(structure_errors) == 0 or \
        all(err['code'] == 'invalid-column-delimiter' for err in structure_errors)
160

Pierre Dittgen's avatar
Pierre Dittgen committed
161 162 163 164 165 166 167 168
    # Checks if a column comparison is needed
    header_errors = ('missing-headers', 'extra-headers', 'wrong-headers-order')
    structure_errors = [{**err, 'in_column_comp': err['code'] in header_errors} for err in structure_errors]
    report['table']['errors']['structure'] = structure_errors
    column_comparison_needed = any(err['in_column_comp'] == True for err in structure_errors)
    column_comparison_table = []
    if column_comparison_needed:
        column_comparison_table = []
Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
169
        field_names = [f['name'] for f in schema_dict.get('fields', [])]
170
        has_case_errors = False
Pierre Dittgen's avatar
Pierre Dittgen committed
171 172
        for t in itertools.zip_longest(headers, field_names, fillvalue=''):
            status = 'ok' if t[0] == t[1] else 'ko'
173 174
            if not has_case_errors and status == 'ko' and t[0].lower() == t[1].lower():
                has_case_errors = True
Pierre Dittgen's avatar
Pierre Dittgen committed
175
            column_comparison_table.append((*t, status))
176 177 178 179 180
        info = {}
        info['table'] = column_comparison_table
        info['has_missing'] = len(headers) < len(field_names)
        info['has_case_errors'] = has_case_errors
        report['table']['column_comparison_info'] = info
Pierre Dittgen's avatar
Pierre Dittgen committed
181 182
    report['table']['column_comparison_needed'] = column_comparison_needed

Pierre Dittgen's avatar
Pierre Dittgen committed
183 184 185 186
    # Group body errors by row id
    rows = []
    current_row_id = 0
    for err in report['table']['errors']['body']:
187 188
        if not 'row-number' in err:
            print('ERR', err)
Pierre Dittgen's avatar
Pierre Dittgen committed
189 190 191 192 193 194 195 196 197 198 199 200 201
        row_id = err['row-number']
        del err['row-number']
        del err['context']
        if row_id != current_row_id:
            current_row_id = row_id
            rows.append({'row_id': current_row_id, 'errors': {}})

        column_id = err.get('column-number')
        if column_id is not None:
            del err['column-number']
            rows[-1]['errors'][column_id] = err
        else:
            rows[-1]['errors']['row'] = err
Pierre Dittgen's avatar
Pierre Dittgen committed
202
    report['table']['errors']['body_by_rows'] = rows
Pierre Dittgen's avatar
Pierre Dittgen committed
203

204 205 206 207 208 209 210 211 212
    # Sort by error names in statistics
    stats = report['table']['error-stats']
    code_title_map = messages.ERROR_MESSAGE_DEFAULT_TITLE
    for key in ('structure-errors', 'value-errors'):
        # convert dict into tuples with french title instead of error code
        # and sorts by title
        stats[key]['count-by-code'] = sorted([(code_title_map.get(k, k), v) for k, v in stats[key]['count-by-code'].items()],
                                             key=itemgetter(0))

Pierre Dittgen's avatar
Pierre Dittgen committed
213 214 215
    return report


Pierre Dittgen's avatar
Pierre Dittgen committed
216 217 218 219 220 221 222 223 224 225 226 227 228 229 230
def compute_badge_message_and_color(badge):
    """Computes message and color from badge information"""
    structure = badge['structure']
    body = badge.get('body')

    # Bad structure, stop here
    if structure == 'KO':
        return (
            'structure invalide', 'red')

    # No body error
    if body == 'OK':
        return ('structure invalide', 'orange') if structure == 'WARN' else ('valide', 'green')

    # else compute quality ratio percent
Christophe Benz's avatar
Christophe Benz committed
231
    p = (1 - badge['error-ratio']) * 100.0
Pierre Dittgen's avatar
Pierre Dittgen committed
232 233 234 235 236 237 238 239
    msg = 'cellules valides : {:.1f}%'.format(p)
    return (msg, 'red') if body == 'KO' else (msg, 'orange')


def get_badge_url_and_message(badge):
    """Gets badge url from badge information"""

    msg, color = compute_badge_message_and_color(badge)
240 241
    return ('{}static/v1.svg?label=Validata&message={}&color={}'.format(
        config.SHIELDS_IO_BASE_URL, quote_plus(msg), color), msg)
Pierre Dittgen's avatar
Pierre Dittgen committed
242 243


Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
244
def validate(schema_url, source: ValidataSource):
245 246
    """ Validate source and display report """

Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
247 248 249 250 251 252
    if config.API_VALIDATE_ENDPOINT is None:
        flash_error("No Validate endpoint defined :-(")
        return redirect(url_for("custom_validator"))
    api_url = config.API_VALIDATE_ENDPOINT
    headers = {"Accept": "application/json"}

253
    try:
254

Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
255 256 257 258 259 260 261 262 263 264 265 266 267 268 269
        if source.is_url():
            params = {
                "schema": schema_url,
                "url": source.get_url(),
            }
            req = requests.get(api_url, params=params, headers=headers)

        else:
            files = {'file': (source.name, io.BytesIO(source.source))}
            data = {'schema': schema_url}
            req = requests.post(api_url, data=data, files=files, headers=headers)

        if not req.ok:
            flash_error("{}: :(".format(req.status_code))
            return redirect(url_for("home"))
270

Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
271 272 273
        json_response = req.json()
        validata_core_report = json_response['report']
        schema_dict = json_response['schema']
274

Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
275 276 277 278
    except requests.ConnectionError as err:
        logging.exception(err)
        flash_error(str(err))
        return redirect(url_for('home'))
279

Pierre Dittgen's avatar
Pierre Dittgen committed
280
    # Computes badge from report and badge configuration
281
    badge = compute_badge(validata_core_report, config.BADGE_CONFIG)
Pierre Dittgen's avatar
Pierre Dittgen committed
282 283
    badge_url, badge_msg = get_badge_url_and_message(badge)

284 285 286 287 288 289
    source_errors = [err for err in validata_core_report['tables'][0]['errors'] if err['code'] == 'source-error']
    if source_errors:
        err = source_errors[0]
        msg = "l'encodage du fichier est invalide. Veuillez le corriger" if 'charmap' in err[
            'message'] else err['message']
        flash_error('Erreur de source : {}'.format(msg))
Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
290
        return redirect(url_for('custom_validator'))
291

292
    source_data = extract_source_data(source)
293

Pierre Dittgen's avatar
Pierre Dittgen committed
294
    # handle report date
295
    report_datetime = datetime.fromisoformat(validata_core_report['date']).astimezone()
296

Pierre Dittgen's avatar
Pierre Dittgen committed
297
    # Enhance validata_core_report
Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
298
    validata_report = create_validata_ui_report(validata_core_report, schema_dict)
Pierre Dittgen's avatar
Pierre Dittgen committed
299

Pierre Dittgen's avatar
Pierre Dittgen committed
300
    # Display report to the user
301
    return render_template('validation_report.html', title='Rapport de validation',
Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
302
                           val_info={}, report=validata_report,
303
                           validation_date=report_datetime.strftime('le %d/%m/%Y à %Hh%M'),
304
                           source=source, source_type=source.type, source_data=source_data,
305
                           print_mode=request.args.get('print', 'false') == 'true',
Pierre Dittgen's avatar
Pierre Dittgen committed
306
                           badge_url=badge_url, badge_msg=badge_msg,
307
                           report_str=json.dumps(validata_report, sort_keys=True, indent=2),
Pierre Dittgen's avatar
Pierre Dittgen committed
308
                           breadcrumbs=[{'url': url_for('home'), 'title': 'Accueil'},
Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
309 310
                                        {'url': url_for('custom_validator'),
                                         'title': "Rapport de validation"}])
311 312


313 314
def bytes_data(f):
    """ Gets bytes data from Werkzeug FileStorage instance """
Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
315
    iob = io.BytesIO()
316 317 318 319 320
    f.save(iob)
    iob.seek(0)
    return iob.getvalue()


321 322 323 324 325 326
# Routes


@app.route('/')
def home():
    """ Home page """
Pierre Dittgen's avatar
Pierre Dittgen committed
327
    validators = ValidatorHelper.schema_info_list()
Christophe Benz's avatar
Christophe Benz committed
328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344
    external_validators = [
        {
            "title": "INSPIRE",
            "description": "proposé par la Commission Européenne pour tester des ressources géographiques (données, services ou métadonnées)",
            "url": "http://inspire-sandbox.jrc.ec.europa.eu/validator/"
        },
        {
            "title": "BAL",
            "description": "proposé par la mission Etalab pour tester des données Adresse produites localement (format BAL 1.1)",
            "url": "https://adresse.data.gouv.fr/bases-locales/validateur"
        },
        {
            "title": "CVDTC",
            "description": "proposé par l'AFIMB dans le cadre du projet Chouette pour tester des données de transport collectif (GTFS ou NEPTUNE)",
            "url": "http://www.conversion-validation-donnees-tc.org/"
        },
    ]
Christophe Benz's avatar
Christophe Benz committed
345
    flash_warning('Ce service est fourni en mode beta - certains problèmes peuvent subsister - nous mettons tout en œuvre pour améliorer son fonctionnement en continu.')
Christophe Benz's avatar
Christophe Benz committed
346
    return render_template('home.html', title='Accueil', validators=validators, external_validators=external_validators)
347 348 349 350 351 352 353 354


@app.route('/validators')
def validators():
    """ No validators page """
    return redirect(url_for('home'))


355 356 357
@app.route('/pdf/<val_code>')
def pdf_report(val_code):
    """PDF report generation"""
358
    err_prefix = 'Erreur de génération du rapport PDF'
359 360

    if not ValidatorHelper.schema_exist(val_code):
361 362
        flash_error(err_prefix + ': schéma inconnu')
        return redirect(url_for('scdl_validator', val_code=val_code))
363 364 365

    url_param = request.args.get('url')
    if not url_param:
366 367
        flash_error(err_prefix + ': URL non fournie')
        return redirect(url_for('scdl_validator', val_code=val_code))
368 369 370 371

    validation_url = '{}?input=url&print=true&url={}'.format(url_for('scdl_validator', val_code=val_code, _external=True),
                                                             quote_plus(url_param))

372 373
    with tempfile.NamedTemporaryFile(prefix='validata_{}_report_'.format(val_code), suffix='.pdf') as tmpfile:
        tmp_pdf_report = Path(tmpfile.name)
374 375 376 377 378

    cmd = ['chromium', '--headless', '--disable-gpu',
           '--print-to-pdf={}'.format(str(tmp_pdf_report)), validation_url]
    result = subprocess.run(cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
    if result.returncode != 0:
379 380
        flash_error(err_prefix)
        log.error("Command %r returned an error: %r", cmd, result.stdout.decode('utf-8'))
381 382
        if tmp_pdf_report.exists():
            tmp_pdf_report.unlink()
383
        return redirect(url_for('scdl_validator', val_code=val_code))
384

Pierre Dittgen's avatar
Pierre Dittgen committed
385 386
    pdf_filename = 'Rapport de validation {} {}.pdf'.format(val_code,
                                                            datetime.now().strftime('%d-%m-%Y %Hh%M'))
387
    response = make_response(tmp_pdf_report.read_bytes())
Pierre Dittgen's avatar
Pierre Dittgen committed
388
    response.headers.set('Content-disposition', 'attachment', filename=pdf_filename)
389 390 391 392 393 394 395 396
    response.headers.set('Content-type', 'application/pdf')
    response.headers.set('Content-length', tmp_pdf_report.stat().st_size)

    tmp_pdf_report.unlink()

    return response


Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459
@app.route('/validators/custom', methods=['GET', 'POST'])
def custom_validator():
    """ Validator page """

    if config.API_VALIDATE_ENDPOINT is None:
        flash_error("URL de connexion à l'API non indiquée :-(")
        return redirect(url_for('home'))

    if request.method == 'GET':

        input_param = request.args.get('input')
        url_param = request.args.get("url")
        schema_param = request.args.get("schema")

        # First form display
        if input_param is None:
            return render_template('custom_validation_form.html', title="Schéma personnalisé",
                                   val_info=None,
                                   breadcrumbs=[{'url': url_for('home'), 'title': 'Accueil'}, ])

        # Process URL
        else:
            if url_param is None or url_param == '':
                flash_error("Vous n'avez pas indiqué d'url à valider")
                return redirect(url_for('custom_validator'))
            if schema_param is None or schema_param == '':
                flash_error("Vous n'avez pas indiqué d'url de schéma")
                return redirect(url_for('custom_validator'))
            try:
                return validate(schema_param, ValidataSource('url', url_param, url_param))
            except tabulator.exceptions.FormatError as e:
                flash_error('Erreur : Format de ressource non supporté')
                log.info(e)
                return redirect(url_for('custom_validator'))
            except tabulator.exceptions.HTTPError as e:
                flash_error('Erreur : impossible d\'accéder au fichier source en ligne')
                log.info(e)
                return redirect(url_for('custom_validator'))

    else:  # POST
        input_param = request.form.get('input')
        if input_param is None:
            flash_error('Aucun fichier à valider')
            return redirect(url_for('custom_validator'))

        schema_param = request.form.get('schema')
        if schema_param is None:
            flash_error('Aucun schéma défini')
            return redirect(url_for('custom_validator'))

        # File validation
        if input_param == 'file':
            f = request.files.get('file')
            if f is None:
                flash_warning("Vous n'avez pas indiqué de fichier à valider")
                return redirect(url_for('custom_validator'))

            b_content = bytes_data(f)
            return validate(schema_param, ValidataSource('file', f.filename, b_content))

        return 'Bizarre, vous avez dit bizarre ?'


460 461 462 463
@app.route('/validators/<val_code>', methods=['GET', 'POST'])
def scdl_validator(val_code):
    """ Validator page """

Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
464 465 466
    # if val_code == "custom":
    #     return custom_validator()

467 468 469 470 471 472
    if not ValidatorHelper.schema_exist(val_code):
        flash_error('Validateur [{}] inconnu'.format(val_code))
        return redirect(url_for('home'))

    if request.method == 'GET':

Pierre Dittgen's avatar
Pierre Dittgen committed
473
        val_info = ValidatorHelper.schema_info(val_code)
474 475 476 477
        input_param = request.args.get('input')

        # First form display
        if input_param is None or input_param not in ('url', 'example'):
Pierre Dittgen's avatar
wip  
Pierre Dittgen committed
478
            return render_template('scdl_validation_form.html', title="val_info['title']",
479 480 481 482 483
                                   val_info=val_info,
                                   breadcrumbs=[{'url': url_for('home'), 'title': 'Accueil'}, ])

        # Process URL
        else:
484 485
            url_param = request.args.get('url')
            if url_param is None or url_param == '':
486 487
                flash_error("Vous n'avez pas indiqué d'url à valider")
                return redirect(url_for('scdl_validator', val_code=val_code))
488 489
            try:
                return validate(val_code, ValidataSource('url', url_param, url_param))
490 491 492 493
            except tabulator.exceptions.FormatError as e:
                flash_error('Erreur : Format de ressource non supporté')
                log.info(e)
                return redirect(url_for('scdl_validator', val_code=val_code))
494 495 496 497
            except tabulator.exceptions.HTTPError as e:
                flash_error('Erreur : impossible d\'accéder au fichier source en ligne')
                log.info(e)
                return redirect(url_for('scdl_validator', val_code=val_code))
498 499 500 501

    else:  # POST
        input_param = request.form.get('input')
        if input_param is None:
502
            flash_error('Aucun fichier à valider')
503 504 505 506 507 508 509 510
            return redirect(url_for('scdl_validator', val_code=val_code))

        # File validation
        if input_param == 'file':
            f = request.files.get('file')
            if f is None:
                flash_warning("Vous n'avez pas indiqué de fichier à valider")
                return redirect(url_for('scdl_validator', val_code=val_code))
511

512
            return validate(val_code, ValidataSource('file', f.filename, bytes_data(f)))
513 514

        return 'Bizarre, vous avez dit bizarre ?'