Configuring Settings for a Corpus
For each imported corpus, there is a sub-folder in the corpus data folder (per default
~/.annis/v4/) with the corpus name as folder name.
In addition to the actual annotations, this corpus folder contains a configuration file named
corpus-config.toml with all corpus specific configuration.1
The configuration file is in the human-readable and editable TOML format and can be edited in any text editor.
TOML uses so-called tables to specify sub-sections of the configuration.
Inside each table, normal key value pairs like in the Java Properties file format can be used.
E.g. the following configuration file has the tables/sections
[context] # The default context in the user interface default = 5 # Available context sizes sizes = [1, 2, 5, 10] # Maximal context size in tokens max = 50 [view] page_size = 10
Maximal Context Size, Context Steps and Result Page Sizes
The maximal context size of ±n tokens from each search result (for the KWIC view,
but also for other visualizations) can be set for a corpus using the
max key in the
To configure which steps are actually shown in the front-end (up to the maximum
allowed) add the key
sizes and use an array as value.
By default, the context steps 1, 2, 5 or 10 tokens are available.
The default context selected when the corpus is queried, can be set with the
Setting Default Context and Segmentations
In corpora with multiple segmentations, such as historical corpora with conflicting
diplomatic and normalized word form layers, it is possible to choose the default
segmentation for both search context and the KWIC visualization. To set the relevant
segmentations, use the following settings in the
[context] # The default segmentation used to generate the context segmentation="SEGNAME" [view] # The default segmentation to show in the user interface (e.g. the KWIC) base_text_segmentation="SEGNAME"
For more details on segmentations, see the ANNIS Multiple Segmentation Corpora Guide.
Adding Example Queries
User created example queries are stored in the file
corpus-config.toml as an array of tables.
Each example query starts with the line
[[example_queries]] followed by some key value pairs.
[[example_queries]] query = "\"statisch\"" description = "search for the word \"statisch\"" query_language = "AQL" [[example_queries]] query = "/.*lich/" description = "Search for words ending with \"lich\" (regular expression)" query_language = "AQL" [[example_queries]] query = "\"so\" & \"statisch\" & #1 . #2" description = "search for the word \"so\" followed by the word \"statisch\"" query_language = "AQL"
In ANNIS 3 the corpus specific configuration was distributed over several configuration files (e.g. the
corpus.properties file) and database tables (like the
When importing an relANNIS corpus, the previous corpus configuration files will be mapped to the new format automatically.