Coverage for src/crawler/by_source/bdim_crawler.py: 8%
232 statements
« prev ^ index » next coverage.py v7.15.4, created at 2026-09-23 14:47 +0000
1import re
3import lingua
4import regex
5from bs4 import BeautifulSoup, Tag
6from lingua import LanguageDetectorBuilder
7from ptf.cmds.xml.jats.builder.references import (
8 # ContribAuthor,
9 get_all_authors_xml,
10 get_article_title_xml,
11 get_ext_link_xml,
12 get_publisher_xml,
13 get_source_xml,
14 get_volume_xml,
15 get_year_xml,
16)
17from ptf.cmds.xml.jats.jats_parser import parse_mixed_citation_into_ref
18from ptf.cmds.xml.xml_utils import escape
19from ptf.model_data import (
20 ArticleData,
21 create_abstract,
22 create_articledata,
23 create_contributor,
24 create_extid,
25 create_issuedata,
26)
28from crawler.abstract_crawlers.matching_crawler import MatchingCrawler
29from crawler.utils import add_pdf_link_to_xarticle
32class BdimCrawler(MatchingCrawler):
33 source_name = "Biblioteca Digitale Italiana di Matematica"
34 source_domain = "BDIM"
35 source_website = "http://www.bdim.eu"
36 headers = {"accept_encoding": "utf-8", "cookie": "lingua=en; matematica=tex"}
38 title_corrections = {
39 "RLINA_1965_8_39_5_a17": "Eventi fasici nel midollo spinale quali prove di inibizione presinaptica durante il sonno desincronizzato",
40 "RLINA_1973_8_55_6_a0": "Complementarity between nilpotent selfmappings and periodic autohomeomorphisms.",
41 "RLINA_1973_8_55_6_a2": "Sur une extension du lemme de Green.",
42 "RLINA_1979_8_67_1-2_a6": "On the existence o f an unbounded connected set of solutions for nonlinear equations in Banach spaces.",
43 "RLINA_1972_8_52_2_a5": "Sul carattere proiettivo del rapporto plurisezionale.",
44 "RLINA_1980_8_69_1-2_a6": "A note on a variational formulation of the Einstein equations for thermo-elastic materials.",
45 }
47 issue_href = r"\?id=(?P<col>\w+)(?P<issue>_\d{1,4})"
49 _language_detector_builder = LanguageDetectorBuilder.from_languages(
50 lingua.Language.ENGLISH, lingua.Language.FRENCH, lingua.Language.ITALIAN
51 )
53 def __init__(self, *args, **kwargs):
54 super().__init__(*args, **kwargs)
56 def parse_collection_content(self, content):
57 """
58 Parse the HTML page of Annals of Math and returns a list of xissue.
59 Each xissue has its pid/volume/number/year metadata + its url
60 """
61 soup = BeautifulSoup(content, "html.parser")
62 xissues = []
64 reg_issue = regex.compile(self.issue_href)
66 issue_nodes = []
67 for issue in soup.select("div.listafascicoli a"):
68 href = issue.get("href")
69 if isinstance(href, str) and reg_issue.search(href):
70 issue_nodes.append(issue)
72 for issue_node in issue_nodes:
73 # issue_text = issue_node.get_text()
75 part_issue = issue_node.get("href").split("_")
76 volume = part_issue[-2]
77 number = part_issue[-1]
78 year = part_issue[1]
79 serie = part_issue[2]
80 link = "/item" + issue_node.get("href")
81 xissue = self.create_bdim_xissue(link, serie, volume, number, year)
82 if xissue:
83 xissues.append(xissue)
85 return xissues
87 def get_year(self, year):
88 if "/" in year:
89 year = year.split("/")[0]
91 return year
93 def create_bdim_xissue(self, url, serie, volume, number, dates):
94 year = dates.replace("/", "-")
96 xissue = create_issuedata()
97 xissue.pid = f"{self.collection_id}_{year}_{serie}_{volume}_{number}"
98 xissue.fyear = int(year)
99 xissue.volume = volume
100 xissue.number = number
101 xissue.vseries = serie
102 xissue.url = self.source_website + url
104 return xissue
106 def parse_issue_content(self, content, xissue):
107 soup = BeautifulSoup(content, "html.parser")
108 article_nodes = soup.find_all("div", {"class": "referenza"})
110 for index_article, article_node in enumerate(article_nodes):
111 article_link_node = article_node.find("a", text="full entry")
112 if article_link_node:
113 url = article_link_node.get("href")
114 xarticle = create_articledata()
115 xarticle.pid = "a" + str(index_article)
116 xarticle.url = self.source_website + url
118 xissue.articles.append(xarticle)
120 def parse_article_content(self, content, xissue, xarticle, url):
121 """
122 Parse the content with Beautifulsoup and returns an ArticleData
123 """
124 soup = BeautifulSoup(content, "html.parser")
125 # TITLE
126 title_node = soup.select_one("span.titolo")
127 if not title_node:
128 raise ValueError("Couldn't find article title")
129 xarticle.title_tex = title_node.get_text()
131 # Lang
132 if "(Italian)" in title_node.parent.text:
133 xarticle.lang = "it"
134 elif "(English)" in title_node.parent.text:
135 xarticle.lang = "en"
136 # Authors
137 reg_author_link = regex.compile(r"\?testo=\w+")
138 text_author_bloc = soup.select_one("div.referenza p")
139 if text_author_bloc:
140 for link in text_author_bloc.select("a"):
141 href = link.get("href")
142 if isinstance(href, str) and reg_author_link.search(href):
143 contrib_node = link.select_one("span.autore")
144 if contrib_node is not None:
145 surname_node = link.select_one("span.cognome")
146 firstname_node = link.select_one("span.nome")
147 author = create_contributor(role="author")
149 if surname_node is not None:
150 surname = surname_node.get_text()
151 author["last_name"] = surname
153 if firstname_node is not None:
154 firstname = firstname_node.get_text()
155 author["first_name"] = firstname
157 if not firstname_node or not surname_node:
158 string_name = contrib_node.get_text()
159 author["string_name"] = string_name
161 xarticle.contributors.append(author)
163 # ABSTRACT
164 abstract_section_node = soup.select_one("div.sunto")
165 if abstract_section_node:
166 abstract = str(abstract_section_node.get_text())
168 xarticle.abstracts.append(
169 create_abstract(value_tex=abstract, lang=self.detect_language(abstract))
170 )
172 # PDF
173 pdf_url = soup.find_all("a", text="pdf")
174 if len(pdf_url) > 0:
175 pdf_url = self.source_website + pdf_url[0].get("href")
176 add_pdf_link_to_xarticle(xarticle, pdf_url)
178 # PAGES
179 pages = soup.select_one("span.pagine")
180 if pages:
181 pages_to = re.compile(r"(\(?\d+\)?)?-?(\(?\d+\)?)").search(pages.get_text())
182 if pages_to:
183 parts = pages_to[0].split("-")
184 first_page = parts[0].replace("(", "").replace(")", "")
185 if len(parts) > 1:
186 last_page = parts[1].replace("(", "").replace(")", "")
187 xarticle.lpage = last_page
188 xarticle.fpage = first_page
190 # Biblio
191 bibitems_tags = soup.select("div.biblio div.bibitem")
192 xarticle.bibitems = [self.parse_ref(item) for item in bibitems_tags]
194 # metadata
195 bdim_mr_url = "http://www.ams.org/mathscinet-getitem?mr="
196 bdim_zbl_url = "https://zbmath.org/?q=an:"
197 medata_bloc = soup.select_one("div.referenza > p")
198 if not medata_bloc:
199 raise ValueError("metadata_bloc cannot be found")
201 zbl_id = medata_bloc.select_one(f"a[href^='{bdim_zbl_url}']")
203 if zbl_id:
204 extid_href = zbl_id.get("href")
205 if isinstance(extid_href, str):
206 extid = create_extid("zbl-item-id", extid_href.removeprefix(bdim_zbl_url))
207 xarticle.extids.append(extid)
209 mr_id = medata_bloc.select_one(f"a[href^='{bdim_mr_url}']")
210 if mr_id:
211 extid_href = mr_id.get("href")
212 if isinstance(extid_href, str):
213 extid = create_extid("mr-item-id", extid_href.removeprefix(bdim_mr_url))
214 xarticle.extids.append(extid)
216 if xarticle.pid in self.title_corrections:
217 xarticle.title_tex = self.title_corrections[xarticle.pid]
219 content = self.download_file(url, headers={"cookie": "lingua=it; matematica=tex"})
220 xarticle = self.parse_article_content_bdim_it(content, xissue, xarticle, url)
221 return xarticle
223 def parse_article_content_bdim_it(self, content, xissue, xarticle: ArticleData, url):
224 soup = BeautifulSoup(content, "html.parser")
226 # trans abstract
227 abstract_section_node = soup.select_one("div.sunto")
228 if abstract_section_node:
229 abstract = str(abstract_section_node.get_text())
230 if xarticle.abstracts[0]["value_tex"] != abstract:
231 xarticle.abstracts.append(
232 create_abstract(value_tex=abstract, lang=self.detect_language(abstract))
233 )
235 # Trans_title
237 trans_title_node = soup.select_one("span.titolo_trad")
238 if trans_title_node:
239 trans_title = self.create_trans_title(
240 resource_type="article",
241 xresource_lang=xarticle.lang,
242 lang="en" if xarticle.lang == "it" else "it",
243 title_str=trans_title_node.get_text(),
244 )
245 xarticle.titles.append(trans_title)
246 return xarticle
248 def parse_ref(self, item: Tag):
249 value_xml = ""
250 # First pass : we create an semi-complete XML Jats string, except for the authors
251 # that we store inside authors_list to be serialized at the end
252 authors_list = []
253 for c in item.children:
254 c_text = escape(c.text)
255 if isinstance(c, str):
256 value_xml += c_text
257 continue
259 if not isinstance(c, Tag):
260 raise NotImplementedError("bibitem_tag is not a Tag or a string")
262 if c.name == "a":
263 a_xml, is_badge = self.parse_a_tag(c)
264 if is_badge:
265 value_xml = regex.sub(r" \| $", "", value_xml)
266 value_xml += a_xml
267 continue
269 child_class = c.get("class")
270 if not child_class:
271 value_xml += c_text
272 elif "bautore" in child_class:
273 # TODO : parse firstname and lastname
274 author_data, author_xml = self.parse_biblio_author_tag(c, len(authors_list))
275 authors_list.append(author_data)
276 value_xml += author_xml
278 elif "titolo" in child_class:
279 value_xml += get_article_title_xml(c_text)
280 elif "rivista" in child_class:
281 value_xml += get_source_xml(c_text)
282 elif "anno" in child_class:
283 value_xml += get_year_xml(c_text)
284 elif "volume" in child_class:
285 value_xml += get_volume_xml(c_text)
286 elif "publisher" in child_class:
287 value_xml += get_publisher_xml(c_text)
288 else:
289 # booktitle
290 value_xml += c_text
292 # In order to have a valid Jats xml, we have to group all authors into the person-group xml tag.
293 authors_occurence = regex.compile(r"{author_\d}").findall(value_xml)
294 if len(authors_occurence) > 0:
295 first_author = value_xml.index(authors_occurence[0])
296 last_author = value_xml.index(authors_occurence[-1]) + len(authors_occurence[-1])
297 value_xml = (
298 value_xml[:first_author]
299 + get_all_authors_xml(value_xml[first_author:last_author], authors_list)
300 + value_xml[last_author:]
301 )
303 return parse_mixed_citation_into_ref(value_xml)
304 # return self.create_crawled_bibitem([*bib_elements, *bib_link_elements])
306 def parse_a_tag(self, a_tag: Tag):
307 a_text = escape(a_tag.text)
308 href = a_tag.get("href")
309 if not href:
310 return a_text, False
311 elif isinstance(href, list):
312 raise ValueError("a tag has multiple href values !")
313 else:
314 a_type = "uri"
315 if a_text.startswith("MR "):
316 a_type = "mr-item-id"
317 a_text = a_text.removeprefix("MR ")
318 elif a_text.startswith("Zbl "):
319 a_type = "zbl-item-id"
320 a_text = a_text.removeprefix("Zbl ")
321 elif a_text == "fulltext (doi)":
322 a_type = "doi"
323 a_text = a_text.removeprefix("http://dx.doi.org/")
324 return get_ext_link_xml(escape(href), a_text, a_type), a_type != "uri"
326 def parse_biblio_author_tag(self, author_tag: Tag, index: int = 0):
327 value_xml = ""
328 author_data = {"template_str": ""}
329 for c in author_tag.children:
330 c_text = escape(c.text)
331 if isinstance(c, str):
332 author_data["template_str"] += c_text
333 continue
335 if not isinstance(c, Tag):
336 raise NotImplementedError("author_tag is not a Tag or a string")
337 # given name = cognome = prénom
338 # surname = nome = nom de famille
339 child_class = c.get("class")
340 if not child_class:
341 value_xml += c_text
342 elif "cognome" in child_class:
343 c.replace_with("{given_names}")
344 author_data["given_names"] = c_text
345 author_data["template_str"] += "{given_names}"
346 elif "nome" in child_class:
347 c.replace_with("{surname}")
348 author_data["surname"] = c_text
349 author_data["template_str"] += "{surname}"
350 value_xml += "{author_" + str(index) + "}"
352 return author_data, value_xml