views.py 16.8 KB
Newer Older
1 2 3
"""
    Routes
"""
4
import copy
Pierre Dittgen's avatar
Pierre Dittgen committed
5
import itertools
6
import json
Christophe Benz's avatar
Christophe Benz committed
7
import logging
8
import subprocess
Christophe Benz's avatar
Christophe Benz committed
9
import tempfile
10
from datetime import datetime
11
from io import BytesIO
12
from operator import itemgetter
13 14
from pathlib import Path
from urllib.parse import quote_plus
15

16
from backports.datetime_fromisoformat import MonkeyPatch
17
from flask import make_response, redirect, render_template, request, url_for
18 19

import tabulator
20 21
import validata_core
from commonmark import commonmark
Pierre Dittgen's avatar
Pierre Dittgen committed
22 23
from validata_core import compute_badge, csv_helpers, messages
from validata_core.loaders import custom_loaders
24

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()

Christophe Benz's avatar
Christophe Benz committed
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

46
    options = {}
47
    if source.format == "csv":
48
        options['delimiter'] = csv_helpers.detect_dialect(source.source, format=source.format, scheme=source.scheme,
49
                                                          custom_loaders=custom_loaders).delimiter
50
    with tabulator.Stream(source.source, format=source.format, scheme=source.scheme, custom_loaders=custom_loaders,
51
                          **options) as stream:
52 53
        for row in stream:
            if header is None:
54
                header = ['' if v is None else v for v in row]
55
            else:
56
                rows.append(list(map(stringify, row)))
Pierre Dittgen's avatar
Pierre Dittgen committed
57
                nb_rows += 1
58
    preview_rows_nb = min(preview_rows_nb, nb_rows)
59 60
    return {'header': header,
            'rows_nb': nb_rows,
61 62 63
            'data_rows': rows,
            'preview_rows_nb': preview_rows_nb,
            'preview_rows': rows[:preview_rows_nb]}
64 65


Pierre Dittgen's avatar
Pierre Dittgen committed
66 67
def improve_errors(errors):
    """Add context to errors, converts markdown content to HTML"""
68

Pierre Dittgen's avatar
Pierre Dittgen committed
69 70
    def improve_err(err):
        """Adds context info based on row-nb presence and converts content to HTML"""
71

Pierre Dittgen's avatar
Pierre Dittgen committed
72 73 74 75
        # 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
76

Pierre Dittgen's avatar
Pierre Dittgen committed
77
        # markdown to HTML (with default values for 'title' and 'content')
Pierre Dittgen's avatar
Pierre Dittgen committed
78

Pierre Dittgen's avatar
Pierre Dittgen committed
79 80 81
        # Set default title if no title
        if not 'title' in err:
            update_keys['title'] = '[{}]'.format(err['code'])
Pierre Dittgen's avatar
Pierre Dittgen committed
82

Pierre Dittgen's avatar
Pierre Dittgen committed
83 84 85 86
        # 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
87

Pierre Dittgen's avatar
Pierre Dittgen committed
88 89 90
        # 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
91

Pierre Dittgen's avatar
Pierre Dittgen committed
92 93 94
        # Message content
        md_content = '*content soon available*' if not 'content' in err else err['content']
        update_keys['content'] = commonmark(md_content)
95

Pierre Dittgen's avatar
Pierre Dittgen committed
96
        return {**err, **update_keys}
97

Pierre Dittgen's avatar
Pierre Dittgen committed
98
    return list(map(improve_err, errors))
99 100


Pierre Dittgen's avatar
Pierre Dittgen committed
101
def create_validata_ui_report(validata_core_report, schema):
102 103 104 105 106 107
    """ 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"
108
        - error messages are improved
109
    """
Pierre Dittgen's avatar
Pierre Dittgen committed
110
    report = copy.deepcopy(validata_core_report)
111 112 113 114 115 116 117 118 119

    # 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']
120 121 122
    # 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
123 124 125 126
    # Handy col_count info
    headers = report['table'].get('headers', [])
    report['table']['col_count'] = len(headers)

Pierre Dittgen's avatar
Pierre Dittgen committed
127
    # Computes column info
128
    schema_fields = schema.get('fields', [])
Pierre Dittgen's avatar
Pierre Dittgen committed
129 130 131 132
    fields_dict = {f['name']: (f.get('title', 'titre non défini'), f.get('description', '')) for f in schema_fields}
    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]
133

134
    # Provide better (french) messages
Pierre Dittgen's avatar
Pierre Dittgen committed
135 136
    errors = improve_errors(report['table']['errors'])
    del report['table']['errors']
137

138 139 140 141 142 143 144
    # 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:
145
        if err['tag'] == 'structure':
146 147 148 149
            report['table']['errors']['structure'].append(err)
        else:
            report['table']['errors']['body'].append(err)

Pierre Dittgen's avatar
Pierre Dittgen committed
150
    # Checks if there are structure errors different to invalid-column-delimiter
Pierre Dittgen's avatar
Pierre Dittgen committed
151 152 153
    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)
154

Pierre Dittgen's avatar
Pierre Dittgen committed
155 156 157 158 159 160 161 162 163
    # 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 = []
        field_names = [f['name'] for f in schema_fields]
164
        has_case_errors = False
Pierre Dittgen's avatar
Pierre Dittgen committed
165 166
        for t in itertools.zip_longest(headers, field_names, fillvalue=''):
            status = 'ok' if t[0] == t[1] else 'ko'
167 168
            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
169
            column_comparison_table.append((*t, status))
170 171 172 173 174
        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
175 176
    report['table']['column_comparison_needed'] = column_comparison_needed

Pierre Dittgen's avatar
Pierre Dittgen committed
177 178 179 180
    # Group body errors by row id
    rows = []
    current_row_id = 0
    for err in report['table']['errors']['body']:
181 182
        if not 'row-number' in err:
            print('ERR', err)
Pierre Dittgen's avatar
Pierre Dittgen committed
183 184 185 186 187 188 189 190 191 192 193 194 195
        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
196
    report['table']['errors']['body_by_rows'] = rows
Pierre Dittgen's avatar
Pierre Dittgen committed
197

198 199 200 201 202 203 204 205 206
    # 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
207 208 209
    return report


Pierre Dittgen's avatar
Pierre Dittgen committed
210 211 212 213 214 215 216 217 218 219 220 221 222 223 224
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
225
    p = (1 - badge['error-ratio']) * 100.0
Pierre Dittgen's avatar
Pierre Dittgen committed
226 227 228 229 230 231 232 233
    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)
234 235
    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
236 237


238
def validate(schema_code, source: ValidataSource):
239 240
    """ Validate source and display report """

Pierre Dittgen's avatar
Pierre Dittgen committed
241
    try:
Pierre Dittgen's avatar
Pierre Dittgen committed
242
        validata_core_report = ValidatorHelper.validate(
243 244
            schema_code=schema_code,
            force_strings=True,
245
            **source.get_tabulator_params()
246
        )
247 248 249 250 251 252

        # Validator.validate() doesn't throw FormatError if source.format is None
        # Just do it manually
        if source.format is None:
            raise tabulator.exceptions.FormatError()

Pierre Dittgen's avatar
Pierre Dittgen committed
253 254 255
    except tabulator.exceptions.FormatError:
        flash_error('Erreur : format de fichier non supporté')
        return redirect(url_for('scdl_validator', val_code=schema_code))
256

257 258 259 260
    except validata_core.MissingHeaderError:
        flash_error("Erreur : impossible d'extraire les données d'entête du fichier source")
        return redirect(url_for('scdl_validator', val_code=schema_code))

Pierre Dittgen's avatar
Pierre Dittgen committed
261
    # Computes badge from report and badge configuration
262
    badge = compute_badge(validata_core_report, config.BADGE_CONFIG)
Pierre Dittgen's avatar
Pierre Dittgen committed
263 264
    badge_url, badge_msg = get_badge_url_and_message(badge)

265 266 267 268 269 270 271 272
    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))
        return redirect(url_for('scdl_validator', val_code=schema_code))

273
    source_data = extract_source_data(source)
274

Pierre Dittgen's avatar
Pierre Dittgen committed
275
    # handle report date
276
    report_datetime = datetime.fromisoformat(validata_core_report['date']).astimezone()
277

Pierre Dittgen's avatar
Pierre Dittgen committed
278 279
    # Enhance validata_core_report
    validata_report = create_validata_ui_report(validata_core_report, ValidatorHelper.schema(schema_code).descriptor)
Pierre Dittgen's avatar
Pierre Dittgen committed
280

Pierre Dittgen's avatar
Pierre Dittgen committed
281
    # Display report to the user
Pierre Dittgen's avatar
Pierre Dittgen committed
282
    val_info = ValidatorHelper.schema_info(schema_code)
283
    return render_template('validation_report.html', title='Rapport de validation',
284
                           val_info=ValidatorHelper.schema_info(schema_code), report=validata_report,
285
                           validation_date=report_datetime.strftime('le %d/%m/%Y à %Hh%M'),
286
                           source=source, source_type=source.type, source_data=source_data,
287
                           print_mode=request.args.get('print', 'false') == 'true',
Pierre Dittgen's avatar
Pierre Dittgen committed
288
                           badge_url=badge_url, badge_msg=badge_msg,
289
                           report_str=json.dumps(validata_report, sort_keys=True, indent=2),
Pierre Dittgen's avatar
Pierre Dittgen committed
290
                           breadcrumbs=[{'url': url_for('home'), 'title': 'Accueil'},
Pierre Dittgen's avatar
Pierre Dittgen committed
291 292
                                        {'url': url_for('scdl_validator', val_code=schema_code),
                                         'title': val_info['title']}])
293 294


295 296 297 298 299 300 301 302
def bytes_data(f):
    """ Gets bytes data from Werkzeug FileStorage instance """
    iob = BytesIO()
    f.save(iob)
    iob.seek(0)
    return iob.getvalue()


303 304 305 306 307 308
# Routes


@app.route('/')
def home():
    """ Home page """
Pierre Dittgen's avatar
Pierre Dittgen committed
309
    validators = ValidatorHelper.schema_info_list()
Christophe Benz's avatar
Christophe Benz committed
310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326
    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
327
    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
328
    return render_template('home.html', title='Accueil', validators=validators, external_validators=external_validators)
329 330 331 332 333 334 335 336


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


337 338 339
@app.route('/pdf/<val_code>')
def pdf_report(val_code):
    """PDF report generation"""
340
    err_prefix = 'Erreur de génération du rapport PDF'
341 342

    if not ValidatorHelper.schema_exist(val_code):
343 344
        flash_error(err_prefix + ': schéma inconnu')
        return redirect(url_for('scdl_validator', val_code=val_code))
345 346 347

    url_param = request.args.get('url')
    if not url_param:
348 349
        flash_error(err_prefix + ': URL non fournie')
        return redirect(url_for('scdl_validator', val_code=val_code))
350 351 352 353

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

Christophe Benz's avatar
Christophe Benz committed
354 355
    with tempfile.NamedTemporaryFile(prefix='validata_{}_report_'.format(val_code), suffix='.pdf') as tmpfile:
        tmp_pdf_report = Path(tmpfile.name)
356 357 358 359 360

    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:
361 362
        flash_error(err_prefix)
        log.error("Command %r returned an error: %r", cmd, result.stdout.decode('utf-8'))
363 364
        if tmp_pdf_report.exists():
            tmp_pdf_report.unlink()
365
        return redirect(url_for('scdl_validator', val_code=val_code))
366

Pierre Dittgen's avatar
Pierre Dittgen committed
367 368
    pdf_filename = 'Rapport de validation {} {}.pdf'.format(val_code,
                                                            datetime.now().strftime('%d-%m-%Y %Hh%M'))
Christophe Benz's avatar
Christophe Benz committed
369
    response = make_response(tmp_pdf_report.read_bytes())
Pierre Dittgen's avatar
Pierre Dittgen committed
370
    response.headers.set('Content-disposition', 'attachment', filename=pdf_filename)
371 372 373 374 375 376 377 378
    response.headers.set('Content-type', 'application/pdf')
    response.headers.set('Content-length', tmp_pdf_report.stat().st_size)

    tmp_pdf_report.unlink()

    return response


379 380 381 382 383 384 385 386 387 388
@app.route('/validators/<val_code>', methods=['GET', 'POST'])
def scdl_validator(val_code):
    """ Validator page """

    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
389
        val_info = ValidatorHelper.schema_info(val_code)
390 391 392 393
        input_param = request.args.get('input')

        # First form display
        if input_param is None or input_param not in ('url', 'example'):
394
            return render_template('validation_form.html', title=val_info['title'],
395 396 397 398 399
                                   val_info=val_info,
                                   breadcrumbs=[{'url': url_for('home'), 'title': 'Accueil'}, ])

        # Process URL
        else:
400 401
            url_param = request.args.get('url')
            if url_param is None or url_param == '':
402 403
                flash_error("Vous n'avez pas indiqué d'url à valider")
                return redirect(url_for('scdl_validator', val_code=val_code))
404 405
            try:
                return validate(val_code, ValidataSource('url', url_param, url_param))
406 407 408 409
            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))
410 411 412 413
            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))
414 415 416 417

    else:  # POST
        input_param = request.form.get('input')
        if input_param is None:
Pierre Dittgen's avatar
Pierre Dittgen committed
418
            flash_error('Aucun fichier à valider')
419 420 421 422 423 424 425 426
            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))
427

428
            return validate(val_code, ValidataSource('file', f.filename, bytes_data(f)))
429 430

        return 'Bizarre, vous avez dit bizarre ?'