Add release v1.4 notes

vanderbot/README.md

@@ -32,7 +32,7 @@ Here are some queries that can be run to explore the data:
 
 [Number of clinical trials at Vanderbilt by principal investigator](https://w.wiki/XKK)
 
-The current release is [v1.3](https://github.com/HeardLibrary/linked-data/releases/tag/v1.3).
+The current release is [v1.4](https://github.com/HeardLibrary/linked-data/releases/tag/v1.4).
 
 ## How it works
 
@@ -138,5 +138,11 @@ In the case where there are no reference properties, there also isn't any refere
 If there are reference property combinations other than this, the `generate_statement_data()` function can't be used and custom code must be written for that statement.
 
 
+## Release v1.4 (2020-08-17) notes 
+
+The changes made in this release were made following tests that used the `csv-metadata.json` mapping schema to emit RDF from the source CSV tables. In order to make it possible to create all of the kinds of statements present in the Wikidata data model, the `csv-metadata.json` file and `vb6_upload_wikidata.py` script were changed to use the `ps:` namespace (`http://www.wikidata.org/prop/statement/`) properties rather than the `wdt:` namespace properties. This makes it possible to construct the missing `wdt:` statements using SPARQL CONSTRUCT. [A new script](https://github.com/HeardLibrary/linked-data/blob/master/vanderbot/generate_direct_props.py) materializes those triples by a CONSTRUCT query to a SPARQL endpoint whose triplestore contains the triples generated by the schema. Those materialized triples are then loaded into the triplestore, making it possible to perform queries on any graph pattern that can be used at the Wikidata Query Service SPARQL endpoint.
+
+The first five scripts were not changed in this release.
+
 ----
-Revised 2020-04-23
+Revised 2020-08-17

vanderbot/vb6_upload_wikidata.py

@@ -57,7 +57,11 @@
 # - This requires adding the correct Content-Type header (application/sparql-query)
 # - Correct the form of the IRI for statements (add Q ID before UUID in IRI). This required a slight modification in the 
 #   part of the script that searches the mapping template for statements (look for -} instead of just } )
-
+# -----------------------------------------
+# Version 1.4 change notes (2020-08-17):
+# - In csv-metadata.json, replace wdt: namespace properties with ps: properties, 
+#   e.g. https://github.com/HeardLibrary/linked-data/blob/v1-4/vanderbot/csv-metadata.json#L187
+# - Modify vb6_upload_wikidata.py (this script) to fine those ps: properties instead of the wdt: ones.
 
 import json
 import requests