isidore-en.md 39.6 KB
Newer Older
1
---
2
lang: en
3
description: Presentation of ISIDORE, the search engine for discovering publications, digital data and profiles of social sciences and humanities researchers from around the world.
4
5
6
7
---

# ISIDORE

8
## What is ISIDORE?
9

10
ISIDORE is a search engine for discovering and finding publications, digital data and the profiles of researchers in the social sciences and humanities (SSH) from around the world.
11

12
It allows searching the full text of several million documents (articles, theses and dissertations, reports, datasets, web pages, database records, descriptions of archival holdings, etc.) and event reports (seminars, conferences, etc.). In addition, ISIDORE links these millions of documents together by enriching them with scientific concepts from the work of the SSH research communities.
13

Nicolas Larrousse's avatar
Nicolas Larrousse committed
14
It is accessible on the Web through the portal [isidore.science](https://isidore.science).
15

16
It also offers scientific social network functionalities. As such, it falls into the category of search engines and assistants and offers many features to organize scientific monitoring.
17

18
Launched on December 8, 2010, ISIDORE is the result of a collaboration between the CNRS "very large equipment" Adonis (2007-2013), the Center for Direct Scientific Communication and the companies Antidot, Mondéca and Sword. It is currently developed, updated and operated by the TGIR Huma-Num.
19

20
References on the history of ISIDORE :
21
22

- Yannick Maignien, "ISIDORE, de l'interconnexion de données à l'intégration de services", Hyper Article en Ligne - Sciences de l'Homme et de la Société, [10670/1.k9lck9](https://isidore.science/document/10670/1.k9lck9)
23
- Stéphane Pouyllau et al, "Bilan 2011 de la plateforme ISIDORE et perspectives 2012-2015", MoDyCo, Modèles, Dynamiques, Corpus - UMR 7114, [10670/1.bqexsj](https://isidore.science/document/10670/1.bqexsj)
24
25
- Philippe Bourdenet, "L'espace documentaire en restructuration : l'évolution des services des bibliothèques universitaires", Le serveur TEL (thèses-en-ligne), [10670/1.lnieuv](https://isidore.science/document/10670/1.lnieuv)

26
## How does ISIDORE work?
27

Nicolas Larrousse's avatar
Nicolas Larrousse committed
28
ISIDORE harvests textual metadata and full text, enriches them and then indexes them. It exploits the metadata of the documents as well as the full text, the goal is to analyze this information in order to enrich them, to link them to the concepts of the scientific vocabularies (thesaurus, etc.), to link them to the authors' identifiers (ORCID, IDRef, IDHAL, VIAF, etc.).
29

30
Several enrichments are performed:
31

Nicolas Larrousse's avatar
Nicolas Larrousse committed
32
- Semantic annotation: the words present in the metadata of the documents are compared to the entries of the vocabularies through an algorithm based on a morphological analysis of the terms. If an equivalence is found between a term from the document and an entry in one of the vocabularies, then the resource will be linked to that vocabulary entry. The vocabularies are multilingual and aligned with each other. Thus, the semantic annotation is multilingual.
33

34
- Disciplinary categorization: ISIDORE uses a semantic classifier that, after being trained on a reference corpus, categorizes all documents in ISIDORE into the SSH disciplines of the MORESS repository. The training of the classifier is done with the help of the manual categorization completed by researchers in HAL during the deposit of their publications.
35

36
- The detection of the authors: ISIDORE detects the authors of the documents and enriches the author form (first name and last name) with the help of international (ORCID, VIAF, ISNI) and national (IDHAL, IDRef) author identifiers.
37

38
ISIDORE indexes, in its search engine:
39

40
41
42
43
44
- Document metadata;
- The full text (if it is available in open access) ;
- The semantic annotations ;
- Disciplinary classification;
- Author enrichment and normalization.
45

Nicolas Larrousse's avatar
Nicolas Larrousse committed
46
More information is available on [the "Vocabularies" page](https://isidore.science/vocabularies) of ISIDORE.
47

48
### Can ISIDORE index multilingual documents and data?
49

50
Yes. Since 2015, documents and datasets in English, Spanish
51
and French are indexed, enriched and linked to scientific repositories by ISIDORE (metadata and full text). For full text outside these three languages, it is indexed in the language of the document but in that cas no enrichment is done.
Nicolas Larrousse's avatar
Nicolas Larrousse committed
52
For more information, you can consult our blog post on the subject: [Isidore speaks English, sino también español et toujours en français](https://humanum.hypotheses.org/921).
53

54
### How often is ISIDORE updated?
55

56
57
58
59
60
61
62
ISIDORE is updated, incrementally, on average once a
month. Why this delay? In addition to harvesting and indexing documents
documents, ISIDORE enriches them with concepts from
scientific repositories (thesaurus, taxonomy, etc.). This semantic enrichment work is
automatic and allows us to offer you reading suggestions. It
is to help you discover documents other than those you were
you were looking for. This requires a certain amount of processing and calculation time.
Nicolas Larrousse's avatar
Nicolas Larrousse committed
63
64
The updates of the documents concerning you, which will be thus proposed 
 in your user account as documents to be claimed, will also follow
65
will also follow this monthly update rhythm.
66

67
## How to use ISIDORE?
68

69
ISIDORE offers several tools to search, discover, collect and organize the contents it indexes:
70

71
### The isidore.science portal
72

73
The [isidore.science](https://isidore.science) portal is a website in three languages that provides a [relevance search engine](https://isidore.science) that can be used with several query methods.
74

75
76
77
78
79
80
81
82
- By default, ISIDORE searches for all the words in a query posed by the
    user by removing empty words ("of", "the", "the",
    "the", etc.);
- It is possible to search for a document with a complete sentence or
    It is possible to search for a document with a complete sentence or a group of words by using quotation marks around the sentence,
    for example: "direction of consciousness" will search for exactly
    this sentence. Thus, in this case, the "of" will not be considered
    as an empty word;
83

84
85
86
87
#### Search operators
Several boolean search operators are available in
ISIDORE. Note that the syntax of the operators is important in
ISIDORE, they are always in UPPERCASE (e.g. AND):
88

89
90
91
- AND: the intersection allows to find the terms (or set of terms) common to the query.
    of terms) common to the query. For example:
    - consciousness AND gender
Nicolas Larrousse's avatar
Nicolas Larrousse committed
92
    - "cold war" AND migration
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
- OR: the union allows to find the searched terms
    belonging to both sets of terms, or to one or the other.
    For example:
    - "semantic web" OR "web 3.0"
- EXCEPT (NOT): the exclusion allows to reduce the noise by excluding
    terms. For example:
    - revolution NOT French
- NEAR(n.): the NEAR(n.) operator (understand "close to
    of") allows to link terms by indicating a value "n." of
    proximity between them. It works like an AND with n.
    word(s) between the terms. The value "n." indicates the number of words
    to separate the two terms. NEAR also works
    without the value n. and is in this case equal to a NEAR(10), i.e. 10 words between the
    (10), i.e. 10 words between the searched terms (standard spacing).
    - house NEAR(4) nobility : search for house and nobility with
        proximity of 4 words
109

110
#### Sorting of search results
111

Nicolas Larrousse's avatar
Nicolas Larrousse committed
112
By default, in [isidore.science](https://isidore.science), it is proposed to sort the results by semantic relevance. It is possible to change the sorting of the search results to:
113

Nicolas Larrousse's avatar
Nicolas Larrousse committed
114
- sorting by novelty
115
116
117
- sorting by author's name in alphabetical order
- sorting by author's name in reverse alphabetical order
- sort by ascending date
Nicolas Larrousse's avatar
Nicolas Larrousse committed
118
- sorting by decreasing date
119

Nicolas Larrousse's avatar
Nicolas Larrousse committed
120
Very soon, will be available again:
121

Nicolas Larrousse's avatar
Nicolas Larrousse committed
122
123
- sorting on the title by alphabetical order
- sorting on the title by inverted alphabetical order
124

125
### Advanced Search
126

127
128
An advanced search is also available at [https://isidore.science/as](https://isidore.science/as) and also accessible from
the first page of the [portal](https://isidore.science/as).
129

130
### Personal Space for Researchers
131

Nicolas Larrousse's avatar
Nicolas Larrousse committed
132
Isidore.science offers a personal space for researchers allowing them to:
133

134
135
136
- collect, classify and organize the documents found ;
- gather all their scientific production in order to edit it in a personal profile page;
- follow the productions of colleagues;
137
138
- record and publish queries and their results for monitoring purposes;
- create bibliographies that can be exported to Zotero;
139

140
### The APIs of isidore.science
141

142
143
The [isidore.science search engine APIs](https://api.isidore.science) are available at the URL [https://api.isidore.science](https://api.isidore.science) through the GET method on HTTP or HTTPS.
They provide a fast, accurate and reliable query service for ISIDORE data with advanced search features (auto-completion, spell checking, multi-criteria, boolean and faceted searches, sorting, aggregation of answers, etc).
144

145
Each request to the engine is submitted by means of a URI pointing to a specific web service. The response is a stream in XML (default format) or JSON format.
146

147
The [isidore.science API] web page (https://api.isidore.science) details all the commands available for the different services available.
148

149
### Enriched metadata for *Linked Open Data*.
150

Nicolas Larrousse's avatar
Nicolas Larrousse committed
151
ISIDORE's metadata, ontologies and vocabularies are available in a triplet repository [RDF (Resource Description Framework) or *TripleStore*](https://fr.wikipedia.org/wiki/Resource_Description_Framework), thus placing ISIDORE data in the *Linked Open Data*. A web interface for querying using the SPARQL language and browsing the ISIDORE graph is available via :
152

153
154
- A documented SPARQL query interface and presentation of the ISIDORE data model: https://isidore.science/sqe  
- The basic Virtuoso software interface: https://isidore.science/sparql
155

156
In the ISIDORE *TripleStore*, the main vocabularies for structuring information are :
157

158
- RDF and RDFS
159
160
161
162
163
164
165
166
167
- Dublin Core Element Set
- Duclin Core TERMS
- SIOC
- FOAF
- OWL
- SKOS
- ORE
- DBPEDIA

168
(The complete list is available at <https://isidore.science/sparql?nsdecl>)
169
170


171
### Complementarity between ISIDORE and Zotero
172

173
#### Use from ISIDORE of the Zotero connector to feed its bibliographic database
174

175
ISIDORE is compatible with Zotero and allows to import the references of documents on two levels as soon as the user has installed [the Zotero connector](https://www.zotero.org/download/) in his browser:
176

177
178
- On the page listing the results of a search,
- On the page listing the results of a search, In the page displaying a document.
179

180
#### Using the ISIDORE search connector from Zotero
181

182
Zotero (Linux, MacOS, Windows client) allows to use search engines to search or complete bibliographic references directly from the Zotero interface. We propose here two ISIDORE connectors for Zotero allowing to use ISIDORE from author search.
183

Nicolas Larrousse's avatar
Nicolas Larrousse committed
184
Adding ISIDORE to Zotero allows:
185

186
187
- To complete references from a search on the author's name: this is the "ISIDORE, help me find what he/she has published."
- To find documents in which the author is cited: this is the "ISIDORE, what do you have on the author?"
188

189
These [connectors and installation documentation are available on the TGIR Huma-Num GitLab](https://gitlab.huma-num.fr/spouyllau/ISIDORtero).
190

191
### Use of RSS feeds
192

193
ISIDORE can propose its research results in the form of RSS feeds in order to feed scientific monitoring software (including Zotero for example), research notebooks, etc. The RSS feeds created in ISIDORE are updated, like all the contents of the search engine, approximately once a month during the general update of the ISIDORE contents. Thus, it is possible to follow, from Zotero, the update of the ISIDORE documents resulting from the registered queries.
194

195
To do so, access your personal space (login required), and click "My queries" to see your registered queries:
196

197
![My Image](media/isidore.png)
198

199
200
For a registered query, you have to click on the pictogram "Request
RSS feed of the query" available on the right ![My Image](media/isidore-rss-001.png){: style="width:170px"} and to copy the link with ![My Image](media/isidore-requeteRSS.png){: style="width:120px"}.
201

202
The copied link is of the form: `https://isidore.science/feed/lt3913`.
203

204
205
If your browser is equipped with a module for reading RSS feeds, it will be possible to
possible to use this link directly in your browser.
206
For our example, we will continue with Zotero.
207

208
In Zotero, you have to choose: New feed > From URI :
209

210
![My Image](media/zot-001.png){: style="width:60%;margin-left:20%"}
211

212
Then add the url of the feed provided by ISIDORE (N.B. When using Safari under MacOS, take care to remove the mention "feed:" from
213
214
the url). Then paste it in "URL" of the Zotero RSS feed creation window
of Zotero's RSS feed creation window, example below :
215

216
![My Image](media/zot-002.png)
217

218
Then you have to give a title to your feed, for example :
219
"isidore.science - Query on ...".
220

221
## What can be found in ISIDORE?
222

223
### Organization of documents and data in ISIDORE
224

Nicolas Larrousse's avatar
Nicolas Larrousse committed
225
ISIDORE contains several million documents in SSH that are harvested, enriched with scientific references and indexed. They are organized in :
226

227
228
229
- Research documents and data (archives, raw materials, photographs, films, datasets, statistics, etc.) and are identified in the ISIDORE ontology by: http://isidore.science/class/primaires
- Published documents and data (articles, books, dissertations and theses, reports, etc.) and are identified in the ISIDORE ontology by: http://isidore.science/class/secondaires
- Scientific events (conferences, study days, etc.) and are identified in the ISIDORE ontology by: http://isidore.science/class/evenementielles
230
231


232
233
For a large number of SHS disciplines, ISIDORE allows searching documents coming from the main publication platforms worldwide, as well as a large number of digitized collections from national, university and
municipal libraries.
234
For advanced search uses, the [ISIDORE advanced search](https://isidore.science/as) offers, for example, the possibility of searching for documents between two dates and by discipline or by collections.
235

Nicolas Larrousse's avatar
Nicolas Larrousse committed
236
The main publication platforms (journals and books) present in ISIDORE are:
237
238
239

- OpenEdition
- Cairn
240
- Perseus
241
242
243
244
245
- Erudit
- Oapen
- Redalyc
- Scielo Books

Nicolas Larrousse's avatar
Nicolas Larrousse committed
246
The complete list of collections containing publications can be obtained by querying [the ISIDORE Triple Store](https://isidore.science/sqe) with the [following] SPARQL(https://isidore.science/sparql?query=SELECT+*+WHERE+%7B%0D%0A%3Fs+rdf%3Atype+%3Chttp%3A%2F%2Fisidore. science%2Fclass%2FCollection%3E.%0D%0A%3Fs+rdf%3Atype+%3Chttp%3A%2F%2Fisidore.science%2Fclass%2Fpublications%3E. %0D%0A%3Fs+dcterms%3Atitle+%3Ftitre%0D%0A%7D+ORDER+BY+ASC%28%3Ftitre%29&format=text%2Fhtml&debug=on&timeout=0) :
247
248
249
250
251

```
SELECT * WHERE {
 ?s rdf:type <http://isidore.science/class/Collection>.
 ?s rdf:type <http://isidore.science/class/publications>.
252
253
 ?s dcterms:title ?title
} ORDER BY ASC(?title)
254
```
Nicolas Larrousse's avatar
Nicolas Larrousse committed
255
The main digital libraries (municipal, national, etc.) present in ISIDORE are:
256
257

- Gallica
258
- Selene
259
260
261
262
263
264
265
266
267
268
- E-rara
- NuBIS
- Octaviana
- Burgerbibliothek
- Berkeley Library Digital Collections
- Argonnaute
- BNE
- Cornell
- Didόmena

Nicolas Larrousse's avatar
Nicolas Larrousse committed
269
The complete list of collections containing archival holdings and book collections can be obtained by querying [the ISIDORE Triple Store](https://isidore.science/sqe) with the [following] SPARQL(https://isidore.science/sparql/? default-graph-uri=&query=SELECT+*+WHERE+%7B%0D%0A%3Fs+rdf%3Atype+%3Chttp%3A%2F%2Fisidore.science%2Fclass%2FCollection%3E.%0D%0A%3Fs+rdf%3Atype+%3Chttp%3A%2F%2Fisidore. science%2Fclass%2Fprimaires%3E.%0D%0A%3Fs+dcterms%3Atitle+%3Ftitre%0D%0A%7D+ORDER+BY+ASC%28%3Ftitre%29&format=text%2Fhtml&timeout=0&debug=on) :
270
271
272
273
274

```
SELECT * WHERE {
 ?s rdf:type <http://isidore.science/class/Collection>.
 ?s rdf:type <http://isidore.science/class/primaires>.
275
276
 ?s dcterms:title ?title
} ORDER BY ASC(?title)
277
278
```

279
### Indexing of the main data platforms in SHS
280

281
ISIDORE harvests and indexes the contents of many SSH data platforms allowing researchers to group all their data in their user profile. We encourage researchers, for their research programs, to use platforms offering open interoperability devices and protocols allowing to present documentary and scientific metadata.
282

283
The main data platforms (sources, archives but also publications) are harvested by ISIDORE.
284

285
The complete list of collections can be obtained by querying [the ISIDORE 3store](https://isidore.science/sqe) with the [following] SPARQL(https://isidore.science/sparql/? default-graph-uri=&query=SELECT+*+WHERE+%7B%0D%0A+%3Fs+rdf%3Atype+%3Chttp%3A%2F%2Fisidore.science%2Fclass%2FCollection%3E. %0D%0A+%3Fs+dcterms%3Atitle+%3Ftitre%0D%0A%7D+ORDER+BY+ASC%28%3Ftitre%29%0D%0A&format=text%2Fhtml&timeout=0&debug=on) :
286
287
288
289

```
SELECT * WHERE {
 ?s rdf:type <http://isidore.science/class/Collection>.
290
291
 ?s dcterms:title ?title
} ORDER BY ASC(?title)
292
293
```

294
Please feel free to report any new source to us.
295

296
#### Can data deposited and documented in NAKALA be referenced by ISIDORE?
297

298
Yes, data deposited and documented in NAKALA can be
Nicolas Larrousse's avatar
Nicolas Larrousse committed
299
accessible in ISIDORE. NAKALA offers as standard the [OAI-PMH] interoperability protocol (https://en.wikipedia.org/wiki/Open_Archives_Initiative_Protocol_for_Metadata_Harvesting) which allows for the harvesting of document metadata, and therefore
300
to reference, enrich and index them by ISIDORE.
301

302
303
304
However, referencing by OAI-PMH harvesting is not
automatic for the moment, in particular to allow users to prepare and organize their data and
data and metadata. To be referenced, simply request by email to be indexed ISIDORE via <isidore-sources@huma-num.fr>.
305

306
#### How will scientific articles and images deposited in the HAL, HAL-SHS and MédiHAL open archive be accessible in ISIDORE?
307

308
All the files (PDF, illustrations, photographs, audio and video) deposited and documented in the open archive HAL, including HAL-SHS, as well as MédiHAL are automatically referenced in ISIDORE and indexed at the level of their metadata. All these documents and their notices are thus accessible through the various interfaces of interrogation of ISIDORE.
309

310
#### Can the data deposited in the Didómena (EHESS) warehouse be referenced by ISIDORE?
311

312
Yes, [Didómena](https://didomena.ehess.fr) (the research data warehouse of EHESS) offers OAI-PMH interoperability. Be careful, harvesting is not automatic. To be referenced at the level of your collection, please provide us with the OAI-PMH access point via <isidore-sources@huma-num.fr>.
313

314
#### Can data deposited in Calames (ABES) be referenced by ISIDORE?
315

316
Yes, descriptions of archival holdings cataloged in [Calames](http://calames.abes.fr) (the catalog of archives and manuscripts of French university libraries) are indexed in ISIDORE. However, the EAD-XML standard, used in Calames, does not always allow an optimal documentary indexing: mainly at the level of the richness of the metadata. This is due to the logic of the EAD-XML standard in the encoding of information in the levels of description of the fonds.
317

318
#### Can the data deposited in the Data.sciencespo warehouse be referenced by ISIDORE?
319

320
Yes, the data deposited and documented in [Data.sciencespo](https://data.sciencespo.fr) (Dataverse) offers interoperability in OAI-PMH. It is harvested automatically by ISIDORE.
321

322
#### Can the data deposited in the COCOON platform be referenced by ISIDORE?
323

324
Yes, the data deposited and documented in [the COCOON platform](https://cocoon.huma-num.fr) offers interoperability in OAI-PMH. This platform is automatically harvested by ISIDORE.
325

326
#### Can files and documents deposited in the European Zenodo platform be referenced by ISIDORE?
327

328
329
330
Yes, it is possible for ISIDORE to reference the files and
documents deposited and documented on the platform
Zenodo](https://zenodo.org).
331

332
333
334
335
336
337
338
The referencing is based on the principle of OAI-PMH harvesting on a
set of files and data (and thus their metadata) corresponding to one or more
identifier(s) corresponding to the "communities" identifiers in Zenodo (see https://developers.zenodo.org/#sets).
We can also group several Zenodo identifiers in the same
ISIDORE collection allowing the depositors of several corpora deposited in
deposited in Zenodo to group them in ISIDORE to give them more
visibility.
339

340
341
To add your Zenodo repositories in ISIDORE, [please send us
the URL
342
OAI-PMH](mailto:isidore-sources@huma-num.fr?subject=%22Je%20souhaiterai%20faire%20moissonner%20mes%20dépôts%20Zenodo%22)
343
of your repository (see <https://developers.zenodo.org/#oai-pmh>).
344
345


346
## How do I get data referenced by ISIDORE?
347

348
349
There are several ways to get data and documents referenced by
ISIDORE:
350

351
352
- Submit your data via [an XML stream of standardized metadata and
    using the OAI-PMH protocol](#how-to-signal-data-in-isidore-with-metadata-and-the-oai-pmh-protocol) associated with metadata in
353
    Dublin core format. This method is adapted for documentary databases, corpora, scientific archives and document/data libraries. As an example, [a tool such as
354
355
    such as Omeka (Classic or S) offers the OAI-PMH protocol via modules](#a-website-using-omeka-classic-and-omeka-s-can-be-referenced-by-isidore).
    This method is adapted to research program websites presenting document or data corpora, scientific blogs (except Hypotheses.org), and web pages in general.
356

357
These two methods are also often implemented by data publication tools (CMS, etc.), for example :
358

359
### Can a web site using Drupal be indexed by ISIDORE?
360

361
Yes, it is possible to have web pages generated by the Drupal CMS indexed by ISIDORE.
362
There are two ways to do this, depending on the nature of the
363
content of your pages:
364

365
366
367
368
369
370
371
- Either via the OAI-PMH protocol and in this case there are several
    modules for Drupal, see on
    https://www.drupal.org/search/site/OAI-PMH](https://www.drupal.org/search/site/OAI-PMH?f%5B0%5D=ss_meta_type%3Amodule "OAI-PMH for Drupal").
- Or via the use of a Dublin
    Core metadata structure in the web pages generated by Drupal using RDFa and a
    sitemap.xml. An article dedicated to this way of proceeding is
    available at the above address.
372

373
### Can a website using Omeka Classic and Omeka-S be referenced by ISIDORE?
374

375
Yes, Omeka *Classic* and Omeka S offer modules to expose metadata according to the OAI-PMH protocol:
376

377
378
- Module for [Omeka S](https://omeka.org/s/modules/OaiPmhRepository/)
- Module for [Omeka Classic](https://omeka.org/classic/docs/Plugins/OaiPmhRepository/)
379
380


381
### How to report data in ISIDORE with metadata and OAI-PMH protocol?
382

383
384
To report your data in ISIDORE using the
OAI-PMH protocol, you just have to :
385

386
387
388
- Prepare your data and metadata using the
    Documentary vocabulary Dublin Core Element Set or Dublin Core
    Terms, depending on the level of precision you want, and to
389
    make them accessible via [the OAI-PMH protocol](https://en.wikipedia.org/wiki/Open_Archives_Initiative_Protocol_for_Metadata_Harvesting);
390
- To organize and document the *Sets* in its OAI-PMH repository.
391
- To write to <isidore-sources@huma-num.fr> and give the address of the repository to Huma-Num.
392

393
#### Document sets in OAI-PMH: *Sets*
394

395
The OAI-PMH protocol makes it possible, through the creation of *Sets*, to bring together a coherent set of records whose perimeter makes sense from a scientific or editorial point of view and which is left to the free appreciation of the producer of the data.
396

397
398
399
400
401
402
403
It also allows to define a hierarchy in the *Sets* with an inheritance mechanism by specifying
in the set name the name of the parent *Set* and the child *Set*, separated by the character `:`.
separated by the `:` character. ISIDORE is able to use these
*Sets* to limit harvesting to a set of records or to differentiate between different
differentiate between different data sources within the same warehouse.
The producer will therefore have to specify the harvesting methods that seem to be
appropriate in order to make the most of their resources within ISIDORE.
404
resources within ISIDORE. To do this, he must indicate the *Set* or *Sets*
405
406
concerned or a rule allowing to distinguish the *Sets* to be taken into
to be taken into account.
407

408
The *Sets* can present metadata, in Dublin Core Element Set, which are specific to them. For example:
409

410
``xml
411
412
<set>
 <setSpec>OuvColl</setSpec>
413
 <setName>OuvColl</setName>
414
415
 <setDescription>
  <oai_dc:dc xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:dcterms="http://purl.org/dc/terms/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
416
   <dc:description>Research works distributed on Cairn.info</dc:description>
417
418
419
420
421
  </oai_dc:dc>
 </setDescription>
</set>
```

422
#### Records in OAI-PMH or *Records*:
423

424
425
426
427
428
429
430
In the ISIDORE framework, each OAI-PMH "record" corresponds to a document.
The ISIDORE harvester thus exploits the metadata described according to the
application profile defined by the Open Archive Initiative for the
Dublin Core Element Set (also known as Dublin Core "simple"). Moreover
In addition, the harvester also collects the full-text document(s) whose URLs
whose URLs (beginning with `https://` or `http://`) are specified in the
in the `<dc:identifier>` element.
431

432
433
434
435
We recommend that data producers provide records that are as metadata-rich as possible.
rich as possible in metadata. Indeed, relevance in
ISIDORE favors the richest possible metadata. Fields such as
such as :
436

437
``xml
438
439
440
441
442
<dcterms:description>
<dcterms:creator>
<dcterms:date>
```

443
are essential.
444

445
##### Example of a complete record according to the OAI-PMH protocol:
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461

```xml
<record>
 <header>
  <identifier>oai:halshs.archives-ouvertes.fr:halshs-00514304</identifier>
  <datestamp>2010-09-02T11:06:50Z</datestamp>
  <setSpec>halshs</setSpec>
  <setSpec>SHS:ECO</setSpec>
  <setSpec>SDV:BIO</setSpec>
  <setSpec>INFO:INFO_BT</setSpec>
  <setSpec>SDV:SA:AEP</setSpec>
  <setSpec>SDV:SA:STA</setSpec>
  <setSpec>CIRAD</setSpec>
  <setSpec>SHS</setSpec>
 </header>
 <metadata>
462
  <oai_dc:dc xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
463
464
465
466
467
  <dc:identifier>http://halshs.archives-ouvertes.fr/halshs-00514304/en/ </dc:identifier>
  <dc:identifier>http://halshs.archives-ouvertes.fr/docs/00/51/43/98/PDF/Regulation_GMO_pprint.pdf</dc:identifier>
  <dc:identifier>http://halshs.archives-ouvertes.fr/docs/00/51/43/98/PDF/ppt_nocmt_broader_regulation.pdf </dc:identifier>
  <dc:title>Broadening the scope of regulation: a prerequisite for a positive contribution of transgenic crop useto sustainable development</dc:title>
  <dc:creator>Fok, Michel</dc:creator>
468
  <dc:subject>[SHS:ECO] Humanities and Social Sciences/Economy and finances</dc:subject>
469
470
  <dc:subject>[SDV:BIO] Life Sciences/Biotechnology</dc:subject>
  <dc:subject>[INFO:INFO_BT] Computer Science/Biotechnology</dc:subject>
471
  <dc:subject>[SDV:SA:AEP] Life Sciences/Agricultural sciences/Agriculture, economy and politics</dc:subject>
472
473
474
475
476
477
478
479
480
481
  <dc:subject>[SDV:SA:STA] Life Sciences/Agricultural sciences/Sciences and technics of agriculture</dc:subject>
  <dc:subject>regulation</dc:subject>
  <dc:subject>coordination</dc:subject>
  <dc:subject>GMO</dc:subject>
  <dc:subject>biotechnology</dc:subject>
  <dc:subject>seed price</dc:subject>
  <dc:subject>research</dc:subject>
  <dc:subject>weed resistance</dc:subject>
  <dc:subject>pest complex shift</dc:subject>
  <dc:description>Ex-ante regulation of transgenic crop use generally prevails, before the authorization of commercial release.This kind of regulation addresses the concerns of biosafety and coexistence, under pressure of pros and/or cons of GMO. After fifteen years of large scale use of transgenic crops (notablysoybean and cotton) in various countries (USA, China, Brasil, India...), ecological and economic phenomena are observed and which could threaten the sustainable use of transgenic varieties. I advocate that the regulation scope must be extended so as to a) promote a systemic and coordinatedapproach of transgenic crop use, b) ensure seed purity with regard to the transgenic trait, c) maintain research on non-transgenic varieties, and d) warrant fair pricing of transgenic seeds.</dc:description>
482
  <dc:coverage>Montpelier</dc:coverage>
483
484
485
486
487
488
489
490
491
492
493
  <dc:coverage>France</dc:coverage>
  <dc:date>2010-08-29</dc:date>
  <dc:language>English</dc:language>
  <dc:type>proceeding with peer review</dc:type>
  <dc:source>Proceedings of Agro2010, the XIth ESA Congress</dc:source>
  <dc:source>Agro2010, the XIth ESA Congress</dc:source>
 </oai_dc:dc>
</metadata>
</record>
```

494
495
496
497
In addition to this description in *Dublin Core Element Set*, each
record can be described in one or more metadata formats, the choice of which is
metadata formats, the choice of which is left to the
the administrator of the OAI-PMH warehouse.
498

499
500
The ISIDORE harvester is able to use the *Dublin Core Terms* format and any XML schema allowing
full-text exposure (including TEI or EAD) thus improving its indexing.
501
its indexing. The data producer will have to take care to scrupulously respect the specifications of the OAI-PMH protocol version 2.0, in particular on :
502

503
504
505
- The strict respect of the "datestamp" values in the *records* in order to synchronize the updates between the producer and ISIDORE;
- The good management of deleted data ([detail on the OAI-PMH protocol documentation](http://www.openarchives.org/OAI/openarchivesprotocol.html#DeletedRecords));
- In the case of a publisher's data warehouse or one of significant size, access to its OAI-PMH warehouse via the IP addresses of ISIDORE's OAI-PMH harvesters (harvesting reported by ISIDORE to its IT department).
506

507
508
509
We advise producers to regularly validate the compliance of their warehouse
compliance of their repository using, for example, the [tools of the Open archive
initiative](https://www.openarchives.org/pmh/tools/). Finally, we advise data producers to contact the Huma-Num team for any information requests.
510

511
### How to report data in ISIDORE with RDFa metadata?
512

513
RDFa allows to express a metadata structure according to the principles of the Semantic Web (RDF for *[Resource Description Framework](https://en.wikipedia.org/wiki/Resource_Description_Framework)*) in the HTML code of Web pages. The "a" in RDFa stands for "in
514
attributes", i.e. within the HTML code).
515

516
517
How to express metadata of a web page very simply by
using the [RDFa syntax
518
RDFa](https://tcuvelier.developpez.com/tutoriels/web-semantique/rdfa/introduction/)
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
? For example, in a blog post published with WordPress. If there
may exist [plugins to do this]()
this](https://wordpress.org/plugins/search/RDFa/),
the obsolescence of the latter can make it difficult to maintain them in
over time. Another solution is to implement RDFa in the
HTML code of the WordPress theme you have chosen. For this to be easy
and manageable over time, the easiest way is to use the HTML header
header in order to place `<meta>` tags that will contain some metadata.

Expressing metadata according to the RDF model via the RDFa syntax allows
machines (mainly search engines and indexers) to better process information because it becomes more explicit: for a machine, a string can be a title or a summary, if you don't tell it it's a title or a summary it
will not guess it. So, at the very least, it is possible to use the
tags to define an RDF structure that allows you to structure the minimal metadata
to structure the minimal metadata for example with the Dublin Core Element Set
documentary vocabulary Dublin Core Element Set.

#### How to do it practically?

First of all, it is necessary to indicate in the DOCTYPE of the web page, that it will
contain information that will use the RDF model, so the
DOCTYPE will be :

``xml
542
543
544
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML+RDFa 1.0//EN" "http://www.w3.org/MarkUp/DTD/xhtml-rdfa-1.dtd">
```

545
546
547
In the `<html>` tag, must be present the addresses of the
ontology (via their *NameSpace XML*) which are used to "type
to "type" the information. RDFa - which places metadata in the Semantic Web, requires at least the use of RDF and RDF Schema ontologies and the Dublin Core Element Set (dc). It is possible to use in addition - in order to refine the metadata - the Dublin Core Terms (dcterms):
548

549
``xml
550
551
552
553
554
555
556
<html xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"
xmlns:xsd="http://www.w3.org/2001/XMLSchema#"
xmlns:dc="http://purl.org/dc/elements/1.1/"
xmlns:dcterms="http://purl.org/dc/terms/">
```

557
558
It is possible, to encode more information, to use more
document ontologies:
559

560
``xml
561
562
563
564
565
566
567
568
569
570
571
<html
xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:dc="http://purl.org/dc/elements/1.1/"
xmlns:dcterms="http://purl.org/dc/terms/"
xmlns:skos="http://www.w3.org/2004/02/skos/core#"
xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"
xmlns:xsd="http://www.w3.org/2001/XMLSchema#"
xmlns:foaf="http://xmlns.com/foaf/0.1/"
xmlns:cc="http://creativecommons.org/ns#">
```

572
In the example above, [foaf](http://www.foaf-project.org/) is used to encode information about a person or object described by the metadata. The [CC](https://creativecommons.org) ontology is used to indicate which license, from the *Creative Commons*, would apply to this content.
573

574
575
576
577
The RDFa structure through tags
tags in the `<head>` header of the HTML page. In a first step
first, using a `<link>` tag, we will define the digital object to which the
object to which the RDF encoded information will be attached:
578

579
``xml
580
581
582
<link rel="dc:identifier" href="http://monblog.com/monbillet.html" />
```

583
584
585
586
This tag therefore defines a container for the information that we
This tag defines a container for the information that we are going to indicate using the `<meta>` tags. This container is
identified by a URI which is a URL, i.e. the address of the
the address of the page on the web.
587
588


589
The `<meta>` tags then define a set of metadata, which in our case is descriptive information about the blog post's web page:
590

591
592
593
594
``xml
<meta property="dc:title" content="The title of my post" />
<meta property="dc:creator" content="First name Last name of author 1" />
<meta property="dc:creator" content="First name Last name of author 2" />
595
<meta property="dcterms:created" content="2011-01-27" />
596
<meta property="dcterms:abstract" content="A descriptive summary of my page's content" xml:lang="en" />
597
<meta property="dcterms:abstract" content="A summary in english" xml:lang="en" />
598
599
600
<meta property="dc:subject" content="keyword 3" />
<This is the first time you'll be able to see a summary of the content of the site.
<meta property="dc:type" content="ticket" />
601
<meta property="dc:format" content="text/html" />
602
<meta property="dc:relation" content="A link to a complementary web page" />
603
604
```

605
606
607
608
Depending on the nature of the content of the web page, it is of course possible
to be more precise, more refined and more complete in the information
encoded information. For example, it would be wise to use the DC vocabulary
Terms vocabulary.
609

610
The DC Terms allows for example to include a precise form for a bibliographic reference of the content:
611
612


613
614
``xml
<meta property="dcterms:bibliographicCitation" content="Put a bibliographic reference here" />
615
616
```

617
618
It would be possible to pass the entire text of a web page using the SIOC vocabulary [using the property
property](http://www.lespetitescases.net/rdfaiser-votre-blog-2-la-pratique).
619

620
621
622
It is also possible to link web pages together (to
define a corpus of authors for example) by using in the
DC Terms vocabulary the DC Terms property: `dcterms:isPartOf`.
623
624

```xml
625
<meta property="dcterms:isPartOf" content="URL of another web page" />
626
627
```

628
#### Creating the Sitemap
629

630
631
Once the RDFa encoding is done in the HTML pages, you still need to create
a Sitemap XML file listing the pages you want ISIDORE to harvest and submit the URL of this sitemap:
632

633
``xml
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
	<url>
		<loc>http://monsiteweb.com/</loc>
		<lastmod>2018-01-01</lastmod>
		<changefreq>monthly</changefreq>
		<priority>1.0</priority>
	</url>
	<url>
		<loc>http://monsiteweb.com/page1/</loc>
		<lastmod>2018-03-05</lastmod>
		<changefreq>weekly</changefreq>
		<priority>0.5</priority>
  </url>
</urlset>
```

650
651
652
653
654
655
656
657
658
659
It is possible to test the extraction that ISIDORE will do of your
RDFa metadata using the "ISIDORE on demand" application
available at <https://rd.isidore.science/ondemand/fr/rdfa.html>

## ISIDORE perimeter

### Why are some items not found in ISIDORE?

If you do not find all of your scientific production
in [ISIDORE](https://isidore.science), there may be several
660
explanations. It may be that your articles are published in 
661
journals that are not electronic or that do not make their articles available even
662
 long after they have been published. Indeed, since its
663
664
665
666
667
668
creation, [ISIDORE](https://isidore.science) favors open
access: indexing is better for articles available in
open access. Many electronic journals have made this choice through
through portals such as Open Edition Journal (formerly
Érudit, Persée, and Cairn.info, Redalyc, OApen and the articles of these journals are collected and indexed by
and articles from these journals are therefore collected and indexed by
669
670
[ISIDORE](https://isidore.science).

671
672
673
674
It is also possible that your articles are published online, but not
on an electronic publishing platform (but a website), or on an electronic publishing platform
electronic publishing platform that does not allow indexing via the
(see the question and answer on OAI-PMH).
675

676
677
678
679
680
681
Other journals make their articles available, but only after an
an embargo period. In this case,
[ISIDORE](https://isidore.science) indexes only the metadata
of the article. If you connect via your university library
library, documentation center or via BibCNRS, it is possible that you will still have
you may still have access to these articles.
682

683
684
685
It is possible to search the collections indexed by
ISIDORE](https://isidore.science) by using the engine itself and by
indicating that you want to search the collections.
686

687
688
689
690
It is also possible that your article is published as a PDF image,
in this case only the indexing by
ISIDORE](https://isidore.science) will be allowed, but not its
full text indexing.
691

692
Finally, it is possible that some of your articles are published in
Nicolas Larrousse's avatar
Nicolas Larrousse committed
693
journals that are not classified in SSH.
694

695
696
697
698
In all these cases, you can deposit your articles in an
open archive like HAL (HAL-SHS in particular) which is also indexed by
ISIDORE](https://isidore.science) or contact your
bu/documentation center.
699

700
701
If you are not in any of these cases and thus think that it is an error, you can send us an
error, you can send us an e-mail to isidore@huma-num.fr.
702

703
### Why are some books/chapters of books not reported in ISIDORE?
704

705
706
ISIDORE knows how to identify that a document is of type "book", thus, there are
more than 500,000 books and chapters of books are reported in
707
708
ISIDORE.

709
710
711
It should be noted that there are relatively few platforms for publishing
of online books in open access. ISIDORE indexes in SHS, for example, the
contents of book platforms such as :
712

713
714
715
- [OpenEdition Books](https://isidore.science/search/?collection=10670/3.szxq6s) (at the chapter level, and to flag them) ;
- Scielo Books](https://isidore.science/search/?collection=10670/3.7oraz1) (Brazil);
- [OApen](https://isidore.science/search/?collection=10670/3.pwofj8) (Netherlands);
716
- [Erudit](https://isidore.science/s/collection?q=erudit) (Canada) ;
717
- ...
718

719
720
721
722
In addition, you can, in agreement with your publisher, deposit your work or
book or chapters of books in the open archive
HAL-SHS](https://halshs.archives-ouvertes.fr). It will then be indexed by
ISIDORE within the framework of the indexing of HAL-SHS and recognized as a chapter of work.
723

724
### Why some databases are not reported in ISIDORE?
725

726
Harvesting by ISIDORE requires standardized and normalized metadata exposure (documentary, scientific, etc.) (either using the OAI-PMH protocol or using an XML Sitemap and RDFa metadata, see above).
727

728
If you know of any databases that are not present in ISIDORE, please let us know so that we can check with their publishers/data producers.
729

730
## ISIDORE trainings
731

732
Here we list training courses, functional presentations and online self-training courses on the use of ISIDORE. Do not hesitate to let us know about any training you would like to organize:
733

734
735
- The Urfist Méditerranée proposes a new e-learning training on Isidore](https://urfist.univ-cotedazur.fr/nouvelle-formation-en-ligne-une-initiation-a-isidore/) (March 2021)
- ["Isidore, my personal research assistant"](https://ig.hypotheses.org/2215) by Johanna Daniel (April 2020)