articles.py 11.7 KB
Newer Older
1 2 3 4 5 6 7
# -*- coding: utf-8 -*-
""" harvest_tools.articles

"""
import traceback


8
from automaton import Automaton
9 10
from base import (family_name_fr,
                  format_author_fr,
11
                  learn_my_authors,
12 13 14 15
                  MSG_CRASH,
                  MSG_FIX_ORIGIN,
                  MSG_IN_DB,
                  MSG_LOAD)
16 17 18 19
from invenio_tools import CheckException
from plugin_dbui import get_id, UNDEF_ID


20 21
MSG_NO_EDITOR = "Reject article is not published"
MSG_TRANSFORM_PREPRINT = "Transform the preprint into an article"
22 23


24
class Articles(Automaton):
25
    """Automaton for articles.
26 27 28 29

    """
    def __init__(self, *args, **kwargs):

30
        Automaton.__init__(self, *args, **kwargs)
31 32 33 34

        # the preprint categories
        self.id_preprint = get_id(self.db.categories, code="PRE")

35
    def check_record(self, record):
36
        """Check the content of the article in order to fix non-conformities.
37

38 39
        Args:
            record (RecordPubli): the MARC12 record describing the article.
40

41 42 43
        Returns:
            bool: ``False`` when a non conformity is found and
                can not be corrected.
44 45 46 47 48 49

        """
        if not Automaton.check_record(self, record):
            return False

        if self.dbg:
LE GAC Renaud's avatar
LE GAC Renaud committed
50
            print "check article record"
51 52 53 54 55 56 57 58

        try:
            self.check.clean_erratum(record)

            if not record.is_published():
                self.logs[-1].reject(MSG_NO_EDITOR, record.year())
                return False

59
            self.check.paper_reference(record)
60 61 62 63
            self.check.submitted(record)
            self.check.year(record)

            self.check.format_editor(record)
64 65 66
            self.check.format_authors(record, format_author_fr)

            self.check.get_my_authors(record, family_name_fr)
67 68 69 70 71 72 73 74 75 76 77 78

        except CheckException as e:
            self.logs[-1].reject(e, record.year())
            return False

        except Exception as e:
            self.logs[-1].reject(MSG_CRASH % e, record.year(), translate=False)
            print traceback.format_exc()
            return False

        return True

79 80 81 82 83 84 85 86 87
    def get_record_by_origin(self,
                             id_publisher=None,
                             my_authors=None,
                             oai_url=None,
                             pages=None,
                             publication_url=None,
                             title=None,
                             volume=None,
                             year=None):
88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107
        """Get an existing record using the origin field and its value
        defined in the ``oai_url`` keyword argument.
        The other arguments are used to transform the corresponding preprint
        into an article.

        Args:
            oai_url (unicode): the OAI identifier of the article.
            id_publisher (int): identifier of the publisher in the database.
            my_authors (unicode): authors of my institute separated by a comma.
            pages (unicode):  the page reference.
            publication_url (unicode): the URL of the publications
            title (unicode): the title of the publication.
            volume (unicode): the volume reference.
            year (unicode): the year of publication.

        Returns:
            tuple: ``(id, status)`` which contains the ``id`` of the record.
            It is equal to ``None`` when nothing is found.
            The ``status`` is equal to one when the existing preprint was
            modified into article, zero otherwise
108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138

        """
        if self.dbg:
            print "check existing article by origin"

        db = self.db

        rec_id = get_id(db.publications, origin=oai_url)
        if not rec_id:
            return (None, 0)

        # not a preprint ?
        if db.publications[rec_id].id_categories != self.id_preprint:
            self.logs[-1].idle(MSG_IN_DB, year)
            return (rec_id, 0)

        # transform a preprint into an article
        self.logs[-1].modify(MSG_TRANSFORM_PREPRINT, year)
        if not self.dry_run:
            db.publications[rec_id] = dict(authors_institute=my_authors,
                                           id_categories=self.id_category,
                                           id_publishers=id_publisher,
                                           id_status=UNDEF_ID,
                                           pages=pages,
                                           publication_url=publication_url,
                                           title=title,
                                           volume=volume,
                                           year=year)

        return (rec_id, 1)

139 140 141 142 143 144 145 146 147 148
    def get_record_by_fields(self,
                             id_publisher=None,
                             my_authors=None,
                             oai_url=None,
                             pages=None,
                             publication_url=None,
                             preprint_number=None,
                             title=None,
                             volume=None,
                             year=None):
149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172
        """Get the record matching the input fields

        Note:
            Fix the field origin when a match is found.

        Note:
            Transform a preprint into article.

        Args:
            id_publisher (int): identifier of the publisher in the database.
            my_authors (unicode): authors of my institute separated by a comma.
            oai_url (unicode): the URL defining the OAI.
            pages (unicode):  the page reference.
            publication_url (unicode): the URL of the publications
            preprint_number (unicode): the preprint number
            title (unicode): the title of the publication.
            volume (unicode): the volume reference.
            year (unicode): the year of publication.

        Returns:
            tuple: ``(id, status)`` which contains the ``id`` of the record.
            It is equal to ``None`` when nothing is found.
            The ``status`` is equal to one when the existing preprint was
            modified into article, zero otherwise
173 174 175

        """
        if self.dbg:
176
            print "get existing article by fields"
177 178 179 180 181 182 183 184 185 186 187 188

        db = self.db

        # check against published articles
        rec_id = get_id(db.publications,
                        id_projects=self.id_project,
                        id_publishers=id_publisher,
                        id_teams=self.id_team,
                        pages=pages,
                        volume=volume,
                        year=year)

189
        # fix origin field
190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229
        if rec_id and not db.publications[rec_id].origin:
            if not self.dry_run:
                db.publications[rec_id] = dict(origin=oai_url)

            self.logs[-1].modify(MSG_FIX_ORIGIN, year)
            return (rec_id, 1)

        if rec_id:
            self.logs[-1].idle(MSG_IN_DB, year)
            return (rec_id, 0)

        # check against published preprint
        # a preprint can be identified by its category which is PRE (15)
        rec_id = get_id(db.publications,
                        id_categories=self.id_preprint,
                        id_projects=self.id_project,
                        id_teams=self.id_team,
                        preprint=preprint_number)

        if not rec_id:
            return (None, 0)

        # transform an existing preprint into article
        # institute authors can be missing in the preprint
        # change also the status
        self.logs[-1].modify(MSG_TRANSFORM_PREPRINT, year)

        if not self.dry_run:
            db.publications[rec_id] = dict(authors_institute=my_authors,
                                           id_categories=self.id_category,
                                           id_publishers=id_publisher,
                                           id_status=UNDEF_ID,
                                           pages=pages,
                                           publication_url=publication_url,
                                           title=title,
                                           volume=volume,
                                           year=year)

        return (rec_id, 1)

230 231
    def insert_record(self, record):
        """Insert an article in the database.
232

233 234 235 236 237
        Note:
            The method assumes that erratum are removed.

        Args:
            record (RecordPubli): the MARC12 record describing the article.
238

239 240 241
        Returns:
            int: one when the record is inserted / updated in the database,
                zero otherwise.
242 243 244 245 246 247 248

        """
        db = self.db

        # alias
        editor = record.paper_editor()
        first_author = record.first_author()
249
        my_authors = record.my_authors
250 251 252 253 254 255 256 257 258
        oai_url = record.oai_url()
        pages = record.paper_pages()
        preprint_number = record.preprint_number()
        publication_url = record.paper_url()
        submitted = record.submitted()[0]
        title = record.title()
        volume = record.paper_volume()
        year = record.paper_year()

259 260
        # get the publisher / collaboration identifier
        id_publisher = self.get_create_publisher(editor)
261
        id_collaboration = self.get_create_collaboration(record.collaboration())
262

263 264
        # get already published articles or preprint
        # A preprint is transform into an article.
265 266 267
        #
        # NOTE: The check is performed by origin then by fields.
        # The latter is useful to cover the case where the record
268
        # is entered by hand or by another harvester.
269
        #
LE GAC Renaud's avatar
LE GAC Renaud committed
270 271 272 273 274 275 276 277 278 279 280

        fields = dict(id_publisher=id_publisher,
                      my_authors=my_authors,
                      oai_url=oai_url,
                      pages=pages,
                      publication_url=publication_url,
                      title=title,
                      volume=volume,
                      year=year)

        rec_id, status = self.get_record_by_origin(**fields)
281 282 283
        if rec_id:
            return status

LE GAC Renaud's avatar
LE GAC Renaud committed
284 285 286 287 288 289 290 291 292 293 294
        fields = dict(id_publisher=id_publisher,
                      my_authors=my_authors,
                      oai_url=oai_url,
                      pages=pages,
                      publication_url=publication_url,
                      preprint_number=preprint_number,
                      title=title,
                      volume=volume,
                      year=year)

        rec_id, status = self.get_record_by_fields(**fields)
295 296 297 298 299
        if rec_id:
            return status

        # eventually insert a new articles in the database
        # try to improve the rescue list for CPPM authors
300
        ret = 1
301 302
        if not self.dry_run:

LE GAC Renaud's avatar
LE GAC Renaud committed
303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321
            fields = dict(authors=record.authors(),
                          authors_institute=my_authors,
                          first_author=first_author,
                          id_categories=self.id_category,
                          id_collaborations=id_collaboration,
                          id_projects=self.id_project,
                          id_publishers=id_publisher,
                          id_status=UNDEF_ID,
                          id_teams=self.id_team,
                          origin=oai_url,
                          pages=pages,
                          preprint=preprint_number,
                          publication_url=publication_url,
                          submitted=submitted,
                          title=title,
                          volume=volume,
                          year=year)

            ret = self._insert_in_db(log_year=year, **fields)
322 323 324 325 326 327 328
            if ret == 1:
                learn_my_authors(db,
                                 authors=record.my_authors,
                                 id_project=self.id_project,
                                 id_team=self.id_team,
                                 year=year)

329
        if ret == 1:
330
            self.logs[-1].load(MSG_LOAD, year)
331
            return 1
332

333
        return 0