Skip to content

Latest commit

 

History

History
123 lines (96 loc) · 4.16 KB

configuration-file.md

File metadata and controls

123 lines (96 loc) · 4.16 KB
Raw

The configuration file format

The basics

The configuration file specifies how the tables and columns of the Ditransitives-style Filemaker data bases map to a CLDF data set.

By default the file is called config.json and located in the etc/ folder of a cldfbench project. The location can be customised by moving the file and modifying the python script in the cldfbench accordingly.

The configuration is written in the JSON format, as a key–value map that specifies a number of supported properties, e.g.:

{
    "custom_columns": { ... },
    "colname_maps": { ... },
    "required_columns": { ... },
    "bibtex_map": { ... }
}

Note 1: Not all properties need to be specified in the file. The program will just use default values for any missing property. In fact, the default settings already cover the whole conversion process of a Ditransitives-style data base, so it is only really necessary to add anything here when custom modifications are made to a data set.

The most reliable way to find out about the default values for every property is to check the file src/ditrans2cldf/config.py in the source code. It contains the defaults stored in as Python dictionary called DEFAULT_CONFIG, which mirrors the structure from the JSON data.

Note 2: Make sure there is no comma after the last element of any list or map, otherwise the config file will not be read correctly. Even after years and years of doing this the trailing comma problem never stopped biting me occasionally.

The supported properties

custom_columns

The custom_columns property specifies, which columns should be added to the tables of the resulting CLDF data set. It contains a map of CLDF table names to a list of column names to be added, e.g.:

{
    "custom_columns": {
        "LanguageTable": ["Language_Family"],
        "CodeTable": ["Map_Icon"]
    }
}

colname_maps

The colname_maps property specifies the assignment of columns from Filemaker tables to columns in the CLDF data set. It contains a map of tables to maps of column assignments, e.g.:

{
    "colname_maps": {
        "languages": {
            "Family::name": "Language_Family"
        },
        "lcodes": {
            "mapicon": "Map_Icon"
        }
    }
}

Note that the data layout is different between the Filemaker export and the CLDF data. The Filemaker export does not have a separate code table, while the CLDF data set collapses language and construction parameters into one table. For this reason, the mapping requires the usage of intermediate table names that cover the union of both table sets. Column mapping uses the following table names:

  • languages: Languages
  • lparameters: Language parameters
  • lcodes: Codes of language parameters
  • lvalues: Values of language parameters
  • constructions: Constructions
  • cparameters: Construction parameters
  • ccodes: Codes of construction parameters
  • cvalues: Values of construction parameters
  • examples: Example sentences

required_columns

The required_columns property specifies which columns are considered mandatory for a data set to be considered well-formed. This property is used to find errors and inconsistencies in a specific data set. It contains a map of table names to lists of column names, e.g.:

{
    "required_columns": {
        "languages": ["Language_Family"],
        "lcodes": ["Map_Icon"]
    }
}

Since sanity checking is done before the creation of the CLDF data set, this property uses the same names as the colname_maps property above.

bibtex_map

The bibtex_map specifies how columns from the References table in the Filemaker export are mapped to BibTeX fields in the bibliography of the CLDF data set. It is a map that maps column names to BibTeX field names, e.g.:

{
    "bibtex_map": {
        "Keywords": "keywords",
        "DOI": "doi"
    }
}