Merge pull request #53 from microbiomedata/issue-52-reconcile

Issue 52 reconcile

.gitignore

@@ -1,6 +1,8 @@
 ## NEVER EVER COMMIT THIS
 googlemaps-api-key.txt
 
+bin
+
 cachedir
 examples/outputs
 

Makefile

@@ -1,8 +1,11 @@
 RUN = poetry run
 
+biosample_sqlite_file = ~/biosample_basex_data_good_subset.db
+
 .PHONY: test clean all
 
-all: clean test examples/outputs/report.tsv
+all: clean test examples/outputs/report.tsv assets/bibo_DocumentStatus.tsv rel_to_oxygen_example \
+examples/outputs/non_attribute_metadata_sel_envs_partial.tsv
 
 # ---------------------------------------
 # Test runner
@@ -14,19 +17,24 @@ test:
 
 clean:
 	find examples -name "*report.tsv" -exec rm -rf {} \;
+	rm -rf assets/*bak
+	rm -rf assets/*tsv
+	rm -rf bin/*jar
+	rm -rf downloads/*db
+	rm -rf downloads/*gz
+	rm -rf downloads/*owl
+	rm -rf downloads/*tsv
+	rm -rf downloads/*zip
+	rm -rf examples/outputs/*tsv
 	rm -rf logs/*log
-	rm -rf examples/outputs/*
-
+	rm -rf target/*
 
 examples/outputs/report.tsv: examples/gold.json
 	$(RUN) annotate-sample -R $@ $<
 
 downloads/mixs6_core.tsv:
 	curl -L -s 'https://docs.google.com/spreadsheets/d/1QDeeUcDqXes69Y2RjU2aWgOpCVWo5OVsBX9MKmMqi_o/export?format=tsv&gid=178015749' > $@
 
-
-biosample_sqlite_file = ~/biosample_basex_data_good_subset.db
-
 examples/outputs/non_attribute_metadata_sel_envs_partial.tsv:
 	$(RUN) sqlite_client_cli \
 		--sqlite_path $(biosample_sqlite_file) \
@@ -36,4 +44,16 @@ examples/outputs/non_attribute_metadata_sel_envs_partial.tsv:
 rel_to_oxygen_example: downloads/mixs6_core.tsv
 	$(RUN) rel_to_oxygen_example \
 		--sqlite_path $(biosample_sqlite_file) \
-		--mixs_core_path $<
+		--mixs_core_path $<
+
+
+bin/robot.jar:
+	curl -s https://api.github.com/repos/ontodev/robot/releases/latest  | grep 'browser_download_url.*\.jar"' |  cut -d : -f 2,3 | tr -d \" | wget -O $@ -i -
+
+downloads/bibo.owl:
+	# --location (-L) pursues redirects
+	curl --location https://raw.githubusercontent.com/structureddynamics/Bibliographic-Ontology-BIBO/master/bibo.owl -o $@
+
+assets/bibo_DocumentStatus.tsv: downloads/bibo.owl bin/robot.jar
+	java -jar bin/robot.jar query --input $< --query sparql/bibo_DocumentStatus.sparql $@
+	sed --in-place=.bak 's/^\?//' $@

assets/bibo_DocumentStatus.tsv

@@ -0,0 +1,10 @@
+status_label
+"accepted"
+"draft"
+"forthcoming"
+"legal"
+"non peer reviewed"
+"peer reviewed"
+"published"
+"rejected"
+"unpublished"

assets/bibo_DocumentStatus.tsv.bak

@@ -0,0 +1,10 @@
+?status_label
+"accepted"
+"draft"
+"forthcoming"
+"legal"
+"non peer reviewed"
+"peer reviewed"
+"published"
+"rejected"
+"unpublished"

assets/placeholder.txt

@@ -0,0 +1 @@
+placeholder

downloads/placeholder.txt

@@ -0,0 +1 @@
+placeholder

examples/outputs/placeholder.txt

@@ -0,0 +1 @@
+placeholder

sample_annotator/__init__.py

@@ -1,10 +1,16 @@
 import os
 
-MAIN_MODEL_DIR = os.path.join(os.path.abspath(os.path.dirname(__file__)), 'model')
-MAIN_SCHEMA_DIR = os.path.join(MAIN_MODEL_DIR, 'schema')
-MIXS_SCHEMA = os.path.join(MAIN_SCHEMA_DIR, 'mixs.json')
+MAIN_MODEL_DIR = os.path.join(os.path.abspath(os.path.dirname(__file__)), "model")
+MAIN_SCHEMA_DIR = os.path.join(MAIN_MODEL_DIR, "schema")
+MIXS_SCHEMA = os.path.join(MAIN_SCHEMA_DIR, "mixs.json")
 
 from .sample_annotator import SampleAnnotator
 from .geolocation.geotools import GeoEngine
+
 # , Message
 from .sample_annotator import AnnotationReport
+
+# # this import seemed to succeed a week or so ago
+# # but not now
+# # and may not even be necessary?
+# from report_model import Message

sparql/bibo_DocumentStatus.sparql

@@ -0,0 +1,8 @@
+prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>
+prefix rdf:  <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
+prefix owl:  <http://www.w3.org/2002/07/owl#>
+prefix bibo:  <http://purl.org/ontology/bibo/>
+select distinct (str(?status_label_raw) as ?status_label)
+where { ?s a bibo:DocumentStatus ;
+rdfs:label ?status_label_raw }
+order by ?status_label

sparql/placeholder.txt

@@ -0,0 +1 @@
+placeholder