Batch load and edit formats - NCIEVS/nci-protege5 GitHub Wiki

The format of the files:

1. Tab-delimited (tsv), and the strings are not quoted. Any quotes will be retained, which you probably don’t want unless the quotes are found within the strings themselves.
2. Files are UTF-8 encoded (a basic ascii text file qualifies as utf-8). Any lines with non-UTF8 characters will be discarded.
3. Byte-order-marks (BOM) in Windows unicode/utf-8 files are not being dealt with and break the batch jobs. In Windows, use an editor like Notepad++ to save files as UTF8 w/o a BOM.

The “id” entries in the descriptions below are the fragment identifiers of the IRIs, i.e. whatever follows the last / or # in the IRI.

term	parent-id
Lab Result	C1234
Pizza Order	C456
Holidays	C4321
Lápiz óptico	C654

1. The term cannot contain embedded control characters, or the ‘?’ or ‘!’ characters (see Special Characters in here).
2. The term will be used to create the rdfs:label, the preferred_name, and the fully qualified synonym (with NCI source and PT term type) of the newly-created class.
3. The new class will be assigned a new code/id property, and this code/id will be used for the fragment identifier of the IRI.
4. The parent-id identifies the superclass under which the new class will be treed. This superclass must already exist, i.e. we can’t create a new class in one row and then use it as a superclass in subsequent row, its identifier must be known prior to the run.
   • We don’t support the case where the fragment is generated by taking the term and modifying it, e.g “hello there” -> “hello_there” so that the fragment can be predicted and used elsewhere in the input file as a parent-id.

Format

entity-id	operation	prop-id	value	value
C123	new	P34	Now is the time
C456	modify	P34	Now is the time	For all good women
C567	delete	P52	foo

Example “by code” (i.e. meaningless fragment identifiers in IRIs)

C123	new	P34	Now is the time
C456	modify	P34	Now is the time	For all good women
C567	delete	P52	foo

Example “by name” (i.e. meaningful fragment identifiers in IRIs)

Gene	new	Editor_Note	Now is the time
Gene	modify	Editor_Note	Now is the time	For all good women
Gene	delete	Editor_Note	Now is the time

entity-id	operation	prop-id	value	ann-id	value	ann-id	value	ann-id	value	prop-id	value	ann-id	value
C123	new	P56	foo bar	P101	NCI	P102	PT	P103	FDA
C456	modify	P56	foo bar	P101	NCI	P102	PT	P103	FDA	P56	foo baz	P101	NCI	P102	PT	P103	FDA
C789	delete	P56	foo baz	P101	NCI	P102	PT	P103	FDA

1. Note that for the modify operation, one can tell where the old value ends and new begins as the property id occurs again, .eg. P56 above. The “qualifier” annotations are included to disambiguate properties that might only differ in their qualifiers, i.e. the property values are the same.
2. We call “complex” the properties that have values which are annotated with other properties.

entity-id	operation	parent-id	replacement-id
C123	new	C789
C456	modify	C998	C889
C789	delete	C996

entity-id	operation	id	mod	fill-id	mod	fill-id
C123	new	P56	some	C876
C456	modify	P56	all	C889	some	C889
C567	delete	P67	some	C776