first working prototype

.github/workflows/update_cache.yaml

@@ -6,6 +6,9 @@ on:
   push:
     branches:
       - main
+  pull_request:
+    branches:
+      - main
 
 jobs:
   update-cache:
@@ -19,8 +22,14 @@ jobs:
           git config --global user.name 'L-CAS GitHub'
           git config --global user.email 'marc@hanheide.net'
       - run: |
-          python3 lcas-bib-export-generator.py
-          git add -f *.bib *.rss *.html
-          git commit -m "automated commit"
+          python -m venv .venv
+          source .venv/bin/activate
+          pip install -r requirements.txt
+          python lcas-bib-export-generator.py
+
+      - if: github.event_name != 'pull_request'
+        run: |
+          git add -f *.bib *.json *.html
+          git commit -m "automated commit `date`"
           git push
 

.gitignore

@@ -1,3 +1,6 @@
-*.bib
-*.rss
-wordpress.html
+#*.bib
+#*.rss
+#wordpress.html
+.venv
+*.pyc
+.DS_Store

config.py

@@ -0,0 +1,73 @@
+class Config:
+
+    staff_dict = {
+        'mhanheide': {
+            'sys_id': '9784d77a-95ab-4b79-b2b7-fec86960cadb',
+        },
+        'tduckett': {
+            'sys_id': '94e90db9-b0da-44bd-93bd-cd7649f35ae8',
+        },
+        'esklar': {
+            'sys_id': '5a55374c-44b0-4a2a-946d-7350a68138aa',
+        },
+        'mrai': {
+            'sys_id': '26bc3d5d-a0b2-4a5f-a14b-0314092cfcf9',
+        },
+        'pbosilj': {
+            'sys_id': 'de6e2d62-190c-425a-9066-f8e2c963abb6',
+        },
+        'gdas': {
+            'sys_id': '66512f78-e480-4ec8-9980-7ce1145ae018',
+        },
+        'lguevara': {
+            'sys_id': 'a304664e-6445-4f1e-b8ad-74fa88abc715',
+        },
+        'smaleki': {
+            'sys_id': '9634ee84-d602-47a5-830b-d0e673ffd788',
+        },
+        'malkhafajiy': {
+            'sys_id': 'd533aaa5-bee2-4439-acbe-686567ae5384',
+        },
+        'apolydoros': {
+            'sys_id': '99d1bfa4-2d7f-4a30-9a81-3de5a3a495f2',
+        },
+        'nbellotto': {
+            'sys_id': 'f487270f-7ff0-47eb-b4b6-656cee47efbf',
+        },
+        'pbaxter': {
+            'sys_id': '4e5d4601-928f-429f-a1ea-4e39a555a968',
+        },
+        'gcielniak': {
+            'sys_id': 'e69a17de-0ef1-4771-8004-646967ea1ac9',
+        },
+        'hcuayahuitl': {
+            'sys_id': '6832ced4-f983-44ca-8bc7-36cde054bf0b',
+        },
+        'chfox': {
+            'sys_id': '4311dbb7-1b10-4844-bba9-20f527168e7b',
+        },
+        'sparsons': {
+            'sys_id': 'f6d2a6e0-3f48-40d7-8c0d-15dcefb9bd47',
+        },
+        'spearson': {
+            'sys_id': '6e38837c-9f3d-49e1-ab0c-dbbdf3362853',
+        },
+        'dbochtis': {
+            'sys_id': '04c88bff-e35d-4e6e-8a2f-01f8aaec3c62',
+        },
+        'rpolvara': {
+            'sys_id': 'f65aae87-dc86-4c92-b0ea-ab120ee3a540',
+        },
+        'fdelduchetto': {
+            'sys_id': '9033283c-a405-472e-886a-a4799bf19b1b',
+        },
+        'aklimchik': {
+            'sys_id': 'bd4f1364-7872-4f53-936f-9bdb21e29fed',
+        },
+        'hharman': {
+            'sys_id': '992df11b-7944-4aa7-a123-f63a28942afb',
+        },
+        'tzied': {
+            'sys_id': '5e45b953-1d84-4210-85f7-5b216c28efff',
+        },
+    }

lcas-bib-export-generator-eprints.py

@@ -0,0 +1,119 @@
+#!/usr/bin/env python
+
+from requests import get
+from sys import stderr
+
+# old URL using IDs that is NOT working reliably:
+# http://eprints.lincoln.ac.uk/cgi/search/archive/advanced/export_lirolem_BibTeX.bib?screen=Search&dataset=archive&_action_export=1&output=BibTeX&exp=0%7C1%7C-date%2Fcreators_name%2Ftitle%7Carchive%7C-%7Ccreators_id%3Acreators_id%3AANY%3AIN%3A002146+801704+801872+801929+003092+002165+001928+802799+002604+504752+504299+002325%7Ctype%3Atype%3AANY%3AEQ%3Aarticle+review+book_section+monograph+conference_item+book+thesis+dataset%7C-%7Ceprint_status%3Aeprint_status%3AANY%3AEQ%3Aarchive%7Cmetadata_visibility%3Ametadata_visibility%3AANY%3AEQ%3Ashow&n=
+
+staff = [
+    'Hanheide, Marc',
+    'Duckett, Tom',
+    'Sklar, Elizabeth',
+    'Saaj, Mini',
+    'Elgeneidy',
+    'Esfahani',
+    'Bosilj',
+    'Calisti, Marcello',
+    'Das, Gautham',
+    'Gao, Junfeng',
+    'Guevara, Leonardo',
+    'Maleki, Sepehr',
+    'Al-Khafajiy',
+    'Polydoros', 
+    'Yue, Shigang',
+    'Bellotto, Nicola',
+    'Baxter, Paul',
+    'Cielniak, Grzegorz',
+    'Cuayahuitl, Heriberto',
+    'Fox, Charles',
+    'Parsons, Simon',
+    'Pearson, Simon',
+    'Bochtis',
+    'Polvara, Riccardo',
+    'del Duchetto, Francesco',
+    'Klimchik, Alexandr',
+    'Rai, Mini',
+    'Harman, Helen',
+    'Zied, Tayeb'
+]
+
+# recent
+years = list(range(2012,2025))
+
+def quote_name(n):
+    return '%%22%s%%22' % n.replace(',','%2C').replace(' ','+')
+
+def quote_names(ns):
+    return [quote_name(n) for n in ns]
+
+
+#http://eprints.lincoln.ac.uk/cgi/search/archive/advanced?screen=Search&dataset=archive&_action_search=Search&documents_merge=ALL&documents=&title_merge=ALL&title=&documents.title_merge=ALL&documents.title=&creators_name_merge=ANY&creators_name=Hanheide+Duckett+Saaj+Sklar+Yue+Bellotto+Baxter&creators_id_merge=ALL&creators_id=&abstract_merge=ALL&abstract=&date_online=&date_accepted=&date=2019-&documents.description_merge=ALL&documents.description=&keywords_merge=ALL&keywords=&subjects_merge=ANY&divisions_merge=ANY&editors_name_merge=ALL&editors_name=&refereed=EITHER&publication_merge=ALL&publication=&projects_merge=ALL&projects=&satisfyall=ALL&order=-date%2Fcreators_name%2Ftitle
+#http://eprints.lincoln.ac.uk/cgi/search/archive/advanced/export_lincoln_BibTeX.bib?screen=Search&dataset=archive&_action_export=1&output=BibTeX&exp=0%7C1%7C-date%2Fcreators_name%2Ftitle%7Carchive%7C-%7Ccreators_name%3Acreators_name%3AANY%3AEQ%3AHanheide+Duckett+Saaj+Sklar+Yue+Bellotto+Baxter%7Cdate%3Adate%3AALL%3AEQ%3A2019-%7C-%7Ceprint_status%3Aeprint_status%3AANY%3AEQ%3Aarchive%7Cmetadata_visibility%3Ametadata_visibility%3AANY%3AEQ%3Ashow&n=&cache=11353462
+
+url_pattern='http://eprints.lincoln.ac.uk/cgi/search/archive/advanced/export_lirolem_BibTeX.bib?screen=Search&dataset=archive&_action_export=1&output=BibTeX&exp=0%%7C1%%7C-%%7Ccreators_name%%3Acreators_name%%3AANY%%3AIN%%3A%s%%7Cdate%%3Adate%%3AALL%%3AEQ%%3A%s'
+
+
+'https://eprints.lincoln.ac.uk/cgi/search/archive/advanced/export_lincoln_AllRSS.rss?screen=Search&dataset=archive&_action_export=1&output=AllRSS&exp=0%7C1%7C-date%2Fcreators_name%2Ftitle%7Carchive%7C-%7Ccreators_name%3Acreators_name%3AALL%3AEQ%3AHanheide'
+
+rss_url_pattern='http://eprints.lincoln.ac.uk/cgi/search/archive/advanced/export_lincoln_RSS2.xml?screen=Search&dataset=archive&_action_export=1&output=RSS2&exp=0%%7C1%%7C-%%7Ccreators_name%%3Acreators_name%%3AANY%%3AIN%%3A%s'
+
+
+def highlight_names(names):
+    ret = ''
+    filtered = [s.split(', ')[0] for s in names]
+    return '|'.join(filtered)
+
+shortcode_pattern=(
+    '[bibfilter group="firstauthor" group_order="desc" format="ieee" order=asc limit=1000 '
+    'file="%s" '
+    'timeout=60000 '
+    'highlight="%s" '
+    'sortauthors=1 '
+    'allow="incollection,mastersthesis,article,conference,techreport,inproceedings" '
+    'author="%s"'
+    ']'
+)
+
+def pubs_year_url(year, staff):
+    return url_pattern % ('%2C+'.join(quote_names(staff)), str(year))
+
+def rss_url(staff):
+    return rss_url_pattern % ('%2C+'.join(quote_names(staff)))
+
+def get_file(bibtex_url):
+    return get(bibtex_url, verify=False, timeout=200).text
+
+years.reverse()
+
+with open('wordpress.html','w') as html_file:
+    print('<p>Download the <a href="%s" target="_blank">BibTeX file of all L-CAS publications</a></p>' % (
+        pubs_year_url('', staff)
+    ), file=html_file)
+
+    with open('lcas.bib', 'w') as all_bib:
+        bibtex = get_file(pubs_year_url('', staff))
+        all_bib.write(bibtex)
+
+    for year in years:
+        print("<h2>%s</h2>" % str(year), file=html_file)
+        bibtex_url = pubs_year_url(year, staff)
+        print('generating for year %d using %s' % (year, bibtex_url), file=stderr)
+        print(shortcode_pattern % (
+            bibtex_url,
+            highlight_names(staff),
+            highlight_names(staff)
+        ), file=html_file)
+
+        bibtex = get_file(bibtex_url)
+        with open('%d.bib' % year, 'w') as bibtex_file:
+            bibtex_file.write(bibtex)
+
+
+print('-------------------------------')
+with open('lcas.rss','w') as rss_file:
+    rssdata = get_file(rss_url(staff))
+    rss_file.write(rssdata)
+    print(rss_url(staff))
+
+