Filing definition

When indexing a document in the ELO repository, the metadata, full text content, and information from feed posts associated with the document are added to the iSearch database.

During the indexing process, full text and feed content is analyzed in steps based on the language and the result is stored in the corresponding Elasticsearch index. In the case of metadata, text fields are analyzed and the contents of each field are stored unchanged (e.g. as keyword, date, etc.).

This makes it easier for users to find the information that is relevant for their specific use case. They can simultaneously search the full text and feed contents as well as the metadata.

Steps before indexing

The metadata of documents in the ELO repository are available in the SQL database and can be immediately accessed by the Indexserver. The full text information, on the other hand, first needs to be extracted from a document. This task is performed by the ELO Textreader (opens new window) module.

The Textreader fetches the document from the Indexserver, extract its full text contents and sends this back to the Indexserver, which stores it as a separate file in the ELO repository. The Textreader also uses OCR software to extract text contents if required (e.g. in image files).

Regular indexing process

The indexing process is performed by the Indexserver. The Indexserver checks at regular intervals (updater interval can be configured in the ELO iSearch configuration) which documents have been added, modified, or deleted since the last indexing run. These need to be updated in the iSearch indexes.

The Indexserver retrieves the metadata and feed information of the documents from the SQL database. Full text contents (if available) are retrieved from the repository with the ELO Textreader. Once processed, the information is forwarded to the iSearch, i.e. the Elasticsearch instance, which then indexes the field contents. In the process, the full text and feed content is analyzed in accordance with the language as described above.

Please note

You should consider carefully which documents to add to the iSearch database. Adding too much text information will reduce search engine performance. It can make sense to exclude certain types of documents (e.g. entire book, brochures, etc.)