isidore-en.md 39.7 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
The full text of several million documents (articles, theses and dissertations, reports, datasets, web pages, database records, descriptions of archival holdings, etc.) and event announcements (seminars, conferences, etc.) can be searched. In addition, ISIDORE links these millions of documents together by enriching them with scientific concepts created by 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

28
ISIDORE harvests textual metadata and full text, enriches them and then indexes them. It uses the metadata of the documents as well as the full text. The goal is to analyze this information in order to enrich the document, to link them to the concepts of the scientific vocabularies (thesaurus, etc.) and 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 vocabulary. The classifier is trained with the help of the manual categorization completed by researchers in [HAL](https://hal.archives-ouvertes.fr/) when depositing their publications.
35

36
- 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
ISIDORE is updated, incrementally, on average once a month. Why this delay? In addition to harvesting and indexing documents, ISIDORE enriches them with concepts from scientific repositories (thesauri, taxonomies, etc.). This semantic enrichment is automatic and allows us to offer you suggestions for reading to help you discover documents other than those you were looking for. This requires a certain amount of processing and calculation time.
The updates of the documents with which you are associated, which will thus be proposed in your user account as documents to be claimed, will also be done monthly.
58

59
60
61
62
63
64
65
### What is the circuit for adding collections in ISIDORE?

Two scenarios:

- A research project, a team, a laboratory, a library can propose collections to be harvested by simple e-mail to <isidore-sources@huma-num.fr>. The Huma-Num team studies the request and exchanges with the requester in order to fully understand how the metadata and the data to be indexed are described. Most often, a first harvest and a first indexing and enrichment are carried out so that the requester can see and analyze how their data will be indexed in ISIDORE. Then, the exchanges potentially continue to adjust the indexing process as well as possible.

- The Huma-Num team identifies a data warehouse or a digital library and contacts the data producer or the structure that distributes this data to exchange and propose harvesting and indexing in ISIDORE. A first harvesting and a first indexing and enrichment are carried out so that the requester can see and analyze how their data will be indexed in ISIDORE. Then, the exchanges potentially continue to adjust the indexing process as well as possible.
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
- 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 a group of words by using quotation marks around the sentence or word group, for example: "direction of consciousness" will search for exactly this expression. Thus, in this case, the "of" will not be considered as an empty word;

78

79
80
81
82
#### 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):
83

84
85
- AND: the intersection will find the terms (or set of terms) common to the query.
    For example:
86
    - consciousness AND gender
Nicolas Larrousse's avatar
Nicolas Larrousse committed
87
    - "cold war" AND migration
88
- OR: the union will find the terms belonging to both sets of terms, or to one or the other.
89
90
    For example:
    - "semantic web" OR "web 3.0"
91
- EXCEPT (NOT): the exclusion will reduce the noise by excluding terms. For example:
92
    - revolution NOT French
93
94
95
96
- NEAR(n.): the NEAR(n.) operator (i.e. "close to") will 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 that 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 searched terms (standard spacing).
    - house NEAR(4) nobility : searches for house and nobility with
        a proximity of 4 words

97

98
#### Sorting of search results
99

100
By default, in [isidore.science](https://isidore.science), the results are sorted by semantic relevance. It is possible to change the sorting of the search results to:
101

Nicolas Larrousse's avatar
Nicolas Larrousse committed
102
- sorting by novelty
103
104
105
- 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
106
- sorting by decreasing date
107

108
VVery soon, two more options will also be available:
109

Nicolas Larrousse's avatar
Nicolas Larrousse committed
110
- sorting on the title by alphabetical order
111
- sorting on the title by reverse alphabetical order
112

113
### Advanced Search
114

115
116
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).
117

Edward Gray's avatar
Edward Gray committed
118
### Personal space for researchers
119

Nicolas Larrousse's avatar
Nicolas Larrousse committed
120
Isidore.science offers a personal space for researchers allowing them to:
121

122
- collect, classify and organize the documents found;
123
124
- gather all their scientific production in order to edit it in a personal profile page;
- follow the productions of colleagues;
125
- record and publish queries and their results for monitoring purposes;
126
- create bibliographies that can be exported to Zotero.
127

128
### The APIs of isidore.science
129

Edward Gray's avatar
Edward Gray committed
130
The [isidore.science search engine APIs](https://api.isidore.science) are available through the GET method on HTTP or HTTPS.
131
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).
132

133
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.
134

Edward Gray's avatar
Edward Gray committed
135
The [isidore.science API web page](https://api.isidore.science) details all the commands available for the different services available.
136

137
### Enriched metadata for *Linked Open Data*.
138

139
ISIDORE's metadata, ontologies and vocabularies are available in a triplet repository [RDF (Resource Description Framework) or *TripleStore*](https://en.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:
140

141
142
- 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
143

144
In the ISIDORE *TripleStore*, the main vocabularies for structuring information are:
145

146
- RDF and RDFS
147
- Dublin Core Element Set
148
- Dublin Core TERMS
149
150
151
152
153
154
155
- SIOC
- FOAF
- OWL
- SKOS
- ORE
- DBPEDIA

156
(The complete list is available at <https://isidore.science/sparql?nsdecl>)
157
158


159
### Complementarity between ISIDORE and Zotero
160

161
#### Use from ISIDORE of the Zotero connector to feed its bibliographic database
162

163
ISIDORE is compatible with Zotero. The references of documents can be imported on two levels as soon as the user has installed [the Zotero connector](https://www.zotero.org/download/) in his browser:
164

165
- On the page listing the results of a search,
166
- On the page listing the results of a search, in the page displaying a document.
167

168
#### Using the ISIDORE search connector from Zotero
169

170
Zotero (Linux, MacOS, Windows client) uses search engines to search or complete bibliographic references directly from the Zotero interface. We propose here two ISIDORE connectors for Zotero that make it possible to use ISIDORE from author search.
171

172
173
174
175
By adding ISIDORE to Zotero you can:

- complete references from a search on the author's name: this is the "ISIDORE, help me find what he/she has published."
- find documents in which the author is cited: this is the "ISIDORE, what do you have on the author?"
176
177


178
These [connectors and installation documentation are available on the TGIR Huma-Num GitLab](https://gitlab.huma-num.fr/spouyllau/ISIDORtero).
179

180
### Use of RSS feeds
181

182
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.
183

184
To do so, access your personal space (login required), and click "My queries" to see your registered queries:
185

186
![My Image](media/isidore.png)
187

188
For a registered query, you have to click on the pictogram "Request
Edward Gray's avatar
Edward Gray committed
189
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"}.
190

191
The copied link is in the form: `https://isidore.science/feed/lt3913`.
192

193
If your browser is equipped with a module for reading RSS feeds, this link can be used directly in your browser.
194
For our example, we will continue with Zotero.
195

196
197
In Zotero, you have to choose: New feed > From URI:

198

199
![My Image](media/zot-001.png){: style="width:60%;margin-left:20%"}
200

201
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
202
203
the url). Then paste it in "URL" of the Zotero RSS feed creation window, example below:

204

205
![My Image](media/zot-002.png)
206

207
Then you have to give a title to your feed, for example:
208
"isidore.science - Query on ...".
209

210
## What can be found in ISIDORE?
211

212
### Organization of documents and data in ISIDORE
213

214
ISIDORE contains several million documents in SSH that are harvested, enriched with scientific references and indexed. They are organized into:
215

216
217
218
- Research documents and data (archives, raw materials, photographs, films, datasets, statistics, etc.), identified in the ISIDORE ontology by: http://isidore.science/class/primaires
- Published documents and data (articles, books, dissertations and theses, reports, etc.), identified in the ISIDORE ontology by: http://isidore.science/class/secondaires
- Scientific events (conferences, study days, etc.), identified in the ISIDORE ontology by: http://isidore.science/class/evenementielles
219
220


221
For a large number of SSH disciplines, ISIDORE makes it possible to search documents coming from the main publication platforms worldwide, as well as a large number of digitized collections from national, university and
222
municipal libraries.
223

224
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.
225

Nicolas Larrousse's avatar
Nicolas Larrousse committed
226
The main publication platforms (journals and books) present in ISIDORE are:
227
228
229

- OpenEdition
- Cairn
230
- Perseus
231
232
233
234
235
- Erudit
- Oapen
- Redalyc
- Scielo Books

Edward Gray's avatar
Edward Gray committed
236
The complete list of collections containing publications can be obtained by querying [the ISIDORE Triple Store](https://isidore.science/sqe) with the [following SPARQL request](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) :
237
238
239
240
241

```
SELECT * WHERE {
 ?s rdf:type <http://isidore.science/class/Collection>.
 ?s rdf:type <http://isidore.science/class/publications>.
242
243
 ?s dcterms:title ?title
} ORDER BY ASC(?title)
244
```
Nicolas Larrousse's avatar
Nicolas Larrousse committed
245
The main digital libraries (municipal, national, etc.) present in ISIDORE are:
246
247

- Gallica
248
- Selene
249
250
251
252
253
254
255
256
257
258
- E-rara
- NuBIS
- Octaviana
- Burgerbibliothek
- Berkeley Library Digital Collections
- Argonnaute
- BNE
- Cornell
- Didόmena

Edward Gray's avatar
Edward Gray committed
259
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 request](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) :
260
261
262
263
264

```
SELECT * WHERE {
 ?s rdf:type <http://isidore.science/class/Collection>.
 ?s rdf:type <http://isidore.science/class/primaires>.
265
266
 ?s dcterms:title ?title
} ORDER BY ASC(?title)
267
268
```

269
### Indexing of the main data platforms in SHS
270

271
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 to present documentary and scientific metadata.
272

273
The main data platforms (sources, archives but also publications) are harvested by ISIDORE.
274

Edward Gray's avatar
Edward Gray committed
275
The complete list of collections can be obtained by querying [the ISIDORE 3store](https://isidore.science/sqe) with the [following SPARQL request](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)
276
277
278
279

```
SELECT * WHERE {
 ?s rdf:type <http://isidore.science/class/Collection>.
280
281
 ?s dcterms:title ?title
} ORDER BY ASC(?title)
282
283
```

284
Please feel free to report any new source to us.
285

286
#### Can data deposited and documented in NAKALA be referenced by ISIDORE?
287

288
Yes, data deposited and documented in [NAKALA (the data repository for SSH by Huma-Num)](https://documentation.huma-num.fr/nakala/) can be
289
290
accessible in ISIDORE. NAKALA offers as standard the [OAI-PMH](https://en.wikipedia.org/wiki/Open_Archives_Initiative_Protocol_for_Metadata_Harvesting) interoperability protocol which allows for the harvesting of document metadata, and therefore for referencing, enrichment and indexation by ISIDORE.

291

292
293
However, referencing by OAI-PMH harvesting is not
automatic for the moment, in particular to allow users to prepare and organize their data and
294
data and metadata. To be referenced, simply request by email to be indexed in ISIDORE via <isidore-sources@huma-num.fr>.
295

296
#### How will scientific articles and images deposited in the HAL, HAL-SHS and MédiHAL open archive be accessible in ISIDORE?
297

298
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 query interfaces of ISIDORE.
299

300
#### Can the data deposited in the Didómena (EHESS) warehouse be referenced by ISIDORE?
301

302
Yes, [Didómena](https://didomena.ehess.fr) (the research data warehouse of EHESS) offers OAI-PMH interoperability. Be careful, harvesting is not automatic. For your collection to be referenced, please provide us with the OAI-PMH access point via <isidore-sources@huma-num.fr>.
303

304
#### Can data deposited in Calames (ABES) be referenced by ISIDORE?
305

306
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: this mainly concerns 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 collections.
307

308
#### Can the data deposited in the Data.sciencespo warehouse be referenced by ISIDORE?
309

310
Yes, the data deposited and documented in [Data.sciencespo](https://data.sciencespo.fr) (Dataverse) offer interoperability in OAI-PMH. They are harvested automatically by ISIDORE.
311

312
#### Can the data deposited in the COCOON platform be referenced by ISIDORE?
313

314
315
Yes, the data deposited and documented in [the COCOON platform](https://cocoon.huma-num.fr) offer interoperability in OAI-PMH. This platform is automatically harvested by ISIDORE.

316

317
#### Can files and documents deposited in the European Zenodo platform be referenced by ISIDORE?
318

319
320
Yes, it is possible for ISIDORE to reference the files and
documents deposited and documented on the platform
321
[Zenodo](https://zenodo.org).
322

323
324
325
326
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
327
ISIDORE collection, allowing the depositors of several corpora
328
329
deposited in Zenodo to group them in ISIDORE to give them more
visibility.
330

331

332
To add your Zenodo repositories in ISIDORE, [please send us the URL OAI-PMH](mailto:isidore-sources@huma-num.fr?subject=%22Je%20souhaiterai%20faire%20moissonner%20mes%20dépôts%20Zenodo%22)
333
of your repository (see <https://developers.zenodo.org/#oai-pmh>).
334
335


336
## How do I get data referenced by ISIDORE?
337

338
339
There are several ways to get data and documents referenced by
ISIDORE:
340

341
- 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 Dublin core format. This method is adapted for documentary databases, corpora, scientific archives and document/data libraries. As an example, [a tool 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).
342
343
This method is adapted to research program websites presenting document or data corpora, scientific blogs (except Hypotheses.org), and web pages in general.

344

345
These two methods are also often implemented by data publication tools (CMS, etc.), for example:
346

347
### Can a web site using Drupal be indexed by ISIDORE?
348

349
Yes, it is possible to have web pages generated by the Drupal CMS indexed by ISIDORE.
350
There are two ways to do this, depending on the nature of the
351
content of your pages:
352

353
- Either via the OAI-PMH protocol and in this case there are several
Edward Gray's avatar
Edward Gray committed
354
    modules for Drupal, see [OAI-PMH for Drupal](https://www.drupal.org/search/site/OAI-PMH?f%5B0%5D=ss_meta_type%3Amodule).
355
356
357
358
- 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.
359

360
### Can a website using Omeka Classic and Omeka-S be referenced by ISIDORE?
361

362
Yes, Omeka *Classic* and Omeka S offer modules to expose metadata according to the OAI-PMH protocol:
363

364
365
- Module for [Omeka S](https://omeka.org/s/modules/OaiPmhRepository/)
- Module for [Omeka Classic](https://omeka.org/classic/docs/Plugins/OaiPmhRepository/)
366
367


368
### How to report data in ISIDORE with metadata and OAI-PMH protocol?
369

370
To report your data in ISIDORE using the
371
OAI-PMH protocol, you just have to:
372

373
374
375
- 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
376
    make them accessible via [the OAI-PMH protocol](https://en.wikipedia.org/wiki/Open_Archives_Initiative_Protocol_for_Metadata_Harvesting);
377
- To organize and document the *Sets* in its OAI-PMH repository.
378
- To write to <isidore-sources@huma-num.fr> and give the address of the repository to Huma-Num.
379

380
#### Document sets in OAI-PMH: *Sets*
381

382
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 discretion of the producer of the data.
383

384
It also makes it possible to define a hierarchy in the *Sets* with an inheritance mechanism by specifying
385
in the set name the name of the parent *Set* and the child *Set*,
386
387
separated by the `:` character. ISIDORE is able to use these
*Sets* to limit harvesting to a set of records or to differentiate between different
388
data sources within the same warehouse.
389
390
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.
391
392
393
394
To do this, he must indicate the *Set* or *Sets*
concerned or a rule enabling the *Sets* to be taken into
account to be distinguished.

395

396
The *Sets* can present metadata, in the Dublin Core Element Set, which are specific to them. For example:
397

Edward Gray's avatar
Edward Gray committed
398
```xml
399
400
<set>
 <setSpec>OuvColl</setSpec>
401
 <setName>OuvColl</setName>
402
403
 <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">
404
   <dc:description>Research works distributed on Cairn.info</dc:description>
405
406
407
408
409
  </oai_dc:dc>
 </setDescription>
</set>
```

410
#### Records in OAI-PMH or *Records*:
411

412
413
414
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
415
Dublin Core Element Set (also known as Dublin Core "simple").
416
In addition, the harvester also collects the full-text document(s) whose URLs
417
418
419
420
421
422
 (beginning with `https://` or `http://`) are specified in the
`<dc:identifier>` element.

We recommend data producers to provide records that are as metadata-rich as possible
 since relevance in
ISIDORE favors the richest possible metadata. Fields such as:
423
424


Edward Gray's avatar
Edward Gray committed
425
```xml
426
427
428
429
430
<dcterms:description>
<dcterms:creator>
<dcterms:date>
```

431
are essential.
432

433
##### Example of a complete record according to the OAI-PMH protocol:
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449

```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>
450
  <oai_dc:dc xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
451
452
453
454
455
  <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>
456
  <dc:subject>[SHS:ECO] Humanities and Social Sciences/Economy and finances</dc:subject>
457
458
  <dc:subject>[SDV:BIO] Life Sciences/Biotechnology</dc:subject>
  <dc:subject>[INFO:INFO_BT] Computer Science/Biotechnology</dc:subject>
459
  <dc:subject>[SDV:SA:AEP] Life Sciences/Agricultural sciences/Agriculture, economy and politics</dc:subject>
460
461
462
463
464
465
466
467
468
469
  <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>
470
  <dc:coverage>Montpelier</dc:coverage>
471
472
473
474
475
476
477
478
479
480
481
  <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>
```

482
483
In addition to this description in the *Dublin Core Element Set*, each
record can be described in one or more metadata formats, the choice of which is left to the
484
the administrator of the OAI-PMH warehouse.
485

486

487
488
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.
489
490
The data producer will have to take care to scrupulously respect the specifications of the OAI-PMH protocol version 2.0, in particular as regards:

491

492
493
494
- 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).
495

496
We advise producers to regularly validate the 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.
497

498
### How to report data in ISIDORE with RDFa metadata?
499

500
RDFa can 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
501
attributes", i.e. within the HTML code).
502

503
How to express metadata of a web page very simply by
504
505
506
using the [RDFa syntax](https://tcuvelier.developpez.com/tutoriels/web-semantique/rdfa/introduction/)
? For example, in a blog post published with WordPress. While there
exist [plugins to do this](https://wordpress.org/plugins/search/RDFa/),
507
the obsolescence of the latter can make it difficult to maintain them
508
509
over time. Another solution is to implement RDFa in the
HTML code of the WordPress theme you have chosen. For this to be easy
510
511
512
and manageable over time, the simplest way is to use the HTML header
in order to place `<meta>` tags that will contain some metadata.

513
514

Expressing metadata according to the RDF model via the RDFa syntax allows
515
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 the machine that it's a title or a summary it
516
517
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
518
519
for example with the Dublin Core Element Set.

520
521
522
523
524

#### 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
525
DOCTYPE will be:
526

Edward Gray's avatar
Edward Gray committed
527
```xml
528
529
530
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML+RDFa 1.0//EN" "http://www.w3.org/MarkUp/DTD/xhtml-rdfa-1.dtd">
```

531
The `<html>` tag must contain the addresses of the
532
ontology (via their *NameSpace XML*) which are used
533
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):
534

535

Edward Gray's avatar
Edward Gray committed
536
```xml
537
538
539
540
541
542
543
<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/">
```

544
To encode more information, It is possible to use more
545
document ontologies:
546

547

Edward Gray's avatar
Edward Gray committed
548
```xml
549
550
551
552
553
554
555
556
557
558
559
<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#">
```

560
IIn 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*, applies to this content.
561

562
The RDFa structure through tags
563
564
565
in the `<head>` header of the HTML page. In a first step
, using a `<link>` tag, we will define the digital object to which the
RDF encoded information will be attached:
566

Edward Gray's avatar
Edward Gray committed
567
```xml
568
569
570
<link rel="dc:identifier" href="http://monblog.com/monbillet.html" />
```

571
572
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
573
page on the web.
574
575


576
The `<meta>` tags then define a set of metadata, which in our case is descriptive information about the blog post's web page:
577

Edward Gray's avatar
Edward Gray committed
578
```xml
579
580
581
<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" />
582
<meta property="dcterms:created" content="2011-01-27" />
583
<meta property="dcterms:abstract" content="A descriptive summary of my page's content" xml:lang="en" />
584
<meta property="dcterms:abstract" content="A summary in english" xml:lang="en" />
585
586
<meta property="dc:subject" content="keyword 3" />
<meta property="dc:type" content="ticket" />
587
<meta property="dc:format" content="text/html" />
588
<meta property="dc:relation" content="A link to a complementary web page" />
589
590
```

591
Depending on the nature of the content of the web page, it is of course possible
592
to be more precise, more refined and more complete in the
593
594
encoded information. For example, it would be wise to use the DC Terms vocabulary.

595

596
The DC Terms allow, for example, a precise form for a bibliographic reference of the content to be included:
597
598


Edward Gray's avatar
Edward Gray committed
599
```xml
600
<meta property="dcterms:bibliographicCitation" content="Put a bibliographic reference here" />
601
602
```

603
It would be possible to describe the entire text of a web page using the SIOC vocabulary [using the
604
property](http://www.lespetitescases.net/rdfaiser-votre-blog-2-la-pratique).
605

606
607
608
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`.
609
610

```xml
611
<meta property="dcterms:isPartOf" content="URL of another web page" />
612
613
```

614
#### Creating the Sitemap
615

616
617
618
Once the RDFa encoding has been done in the HTML pages, you still need to create
a Sitemap XML file listing the pages you want ISIDORE to harvest and to submit the URL of this sitemap:

619

Edward Gray's avatar
Edward Gray committed
620
```xml
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
<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>
```

637
638
639
640
641
642
643
644
645
646
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
647
explanations. It may be that your articles are published in
648
journals that are not electronic or that do not make their articles available even
649
 long after they have been published. Since its
650
creation, [ISIDORE](https://isidore.science) favors open
651
access since indexing is better for articles available in
652
open access. Many electronic journals have made this choice through
653
portals such as Open Edition Journal (formerly Revues.org)
654
Érudit, Persée, and Cairn.info, Redalyc, OApen and
655
and articles from these journals are therefore collected and indexed by
656
657
[ISIDORE](https://isidore.science).

658
659
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
660
that does not allow indexing via the standard protocol
661
(see the question and answer on OAI-PMH).
662

663
Other journals make their articles available, but only after an
664
embargo period. In this case,
665
666
[ISIDORE](https://isidore.science) indexes only the metadata
of the article. If you connect via your university library
667
, documentation center or via BibCNRS,
668
you may still have access to these articles.
669

670
671
The collections indexed by
[ISIDORE](https://isidore.science) can be searched by using the engine itself and by
672
indicating that you want to search the collections.
673

674
It is also possible that your article is published as a PDF image,
675
in which case only the indexing by
Edward Gray's avatar
Edward Gray committed
676
[ISIDORE](https://isidore.science) will be allowed, but not its
677
full text indexing.
678

679
Lastly, it is possible that some of your articles are published in
Nicolas Larrousse's avatar
Nicolas Larrousse committed
680
journals that are not classified in SSH.
681

682
In all these cases, you can deposit your articles in an
683
open archive such as HAL (HAL-SHS in particular) which is also indexed by
Edward Gray's avatar
Edward Gray committed
684
[ISIDORE](https://isidore.science) or contact your
685
bu/documentation center.
686

687
If none of these cases correspond to your problem and you therefore think that there may be an error, you can send us an e-mail to isidore@huma-num.fr.
688

689
### Why are some books/chapters of books not reported in ISIDORE?
690

691
692
ISIDORE knows how to identify that a document is of the type "book", thus, there are
more than 500,000 books and book chapters reported in
693
694
ISIDORE.

695
696
It should be noted that there are relatively few platforms that publish online books in open access. ISIDORE indexes in SSH, for example, the
contents of book platforms such as:
697

698
699

- [OpenEdition Books](https://isidore.science/search/?collection=10670/3.szxq6s) (at the chapter level, and to flag them);
Edward Gray's avatar
Edward Gray committed
700
- [Scielo Books](https://isidore.science/search/?collection=10670/3.7oraz1) (Brazil);
701
- [OApen](https://isidore.science/search/?collection=10670/3.pwofj8) (Netherlands);
702
- [Erudit](https://isidore.science/s/collection?q=erudit) (Canada);
703
- ...
704

705
In addition, you can, in agreement with your publisher, deposit your work or
706
book or book chapters in the open archive
Edward Gray's avatar
Edward Gray committed
707
[HAL-SHS](https://halshs.archives-ouvertes.fr). It will then be indexed by
708
709
ISIDORE within the framework of the indexing of HAL-SHS and recognized as a book chapter.

710

711
### Why are some databases are not reported in ISIDORE?
712

713
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).
714

715
If you know of any databases that are not present in ISIDORE, please inform us so that we can check with their publishers/data producers.
716

717
## ISIDORE training courses
718

719
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 session you would like to organize:
720

721
- The *Urfist Méditerranée* proposes a new e-learning training on Isidore (only in french)](https://urfist.univ-cotedazur.fr/nouvelle-formation-en-ligne-une-initiation-a-isidore/) (March 2021)
722
- ["Isidore, my personal research assistant"](https://ig.hypotheses.org/2215) by Johanna Daniel (April 2020)