Slurp pipeline

- Update: direct_owned_parents methods: Temporarily commented out, pending OAK fixes. Currently doing this in batch using SPARQL for now.
- Add: Param --onto-exclusions-path: Now we no longer consider exluded terms to be possible slurp candidates.
- Bugfix: Labels: Now they are successfully being fetched using OAK.
- Add: New columns to output
- Update: run.sh: Reverted back to what it was before, undoing deletion of --rf, which was done to address pip installs not being persistent within the ODK docker container between runs.

src/ontology/mondo-ingest.Makefile

@@ -344,14 +344,14 @@ slurp/:
 	mkdir -p $@
 
 # min-id: the next available Mondo ID
-# TODO: `pip install` stuff is temporary until ODK docker up to date w/ recent OAK updates
-# TODO: Check if removing --rf from run.sh will fix need for pip install. havent been able to try yet; docker issue
+# todo: `pip install` stuff is temporarily here until we come up with a fix. otherwise docker won't work
 slurp/%.tsv: $(COMPONENTSDIR)/%.owl $(TMPDIR)/mondo.sssom.tsv $(REPORTDIR)/mirror_signature-mondo.tsv | slurp/
-	# pip install --upgrade -r $(RELEASEDIR)/requirements-unlocked.txt
+	pip install --upgrade -r $(RELEASEDIR)/requirements-unlocked.txt
 	python $(SCRIPTSDIR)/migrate.py \
 	--ontology-path $(COMPONENTSDIR)/$*.owl \
 	--sssom-map-path $(TMPDIR)/mondo.sssom.tsv \
 	--onto-config-path metadata/$*.yml \
+	--onto-exclusions-path reports/$*_term_exclusions.txt \
 	--min-id 850000 \
 	--max-id 999999 \
 	--mondo-terms-path $(REPORTDIR)/mirror_signature-mondo.tsv \

src/ontology/run.sh

@@ -45,7 +45,7 @@ if [ -n "$USE_SINGULARITY" ]; then
         -W $WORK_DIR \
         docker://obolibrary/$ODK_IMAGE:$ODK_TAG $TIMECMD "$@"
 else
-    docker run -v $VOLUME_BIND -w $WORK_DIR -e ROBOT_JAVA_ARGS="$ODK_JAVA_OPTS" -e JAVA_OPTS="$ODK_JAVA_OPTS" -ti obolibrary/$ODK_IMAGE:$ODK_TAG $TIMECMD "$@"
+    docker run -v $VOLUME_BIND -w $WORK_DIR -e ROBOT_JAVA_ARGS="$ODK_JAVA_OPTS" -e JAVA_OPTS="$ODK_JAVA_OPTS" --rm -ti obolibrary/$ODK_IMAGE:$ODK_TAG $TIMECMD "$@"
 fi
 
 case "$@" in

src/ontology/slurp/icd10cm.tsv

@@ -1,25 +1,22 @@
-mondo_id	xref	label	definition	parents
-ID	A oboInOwl:hasDbXref	LABEL	A IAO:0000115	SC %
-MONDO:850001	ICD10CM:A00-B99			
-MONDO:850002	ICD10CM:C00-D49			
-MONDO:850003	ICD10CM:D50-D89			
-MONDO:850004	ICD10CM:E00-E89			
-MONDO:850005	ICD10CM:F01-F99			
-MONDO:850006	ICD10CM:G00-G99			
-MONDO:850007	ICD10CM:H00-H59			
-MONDO:850008	ICD10CM:H60-H95			
-MONDO:850009	ICD10CM:I00-I99			
-MONDO:850010	ICD10CM:ICD-10-CM			
-MONDO:850011	ICD10CM:J00-J99			
-MONDO:850012	ICD10CM:K00-K95			
-MONDO:850013	ICD10CM:L00-L99			
-MONDO:850014	ICD10CM:M00-M99			
-MONDO:850015	ICD10CM:N00-N99			
-MONDO:850016	ICD10CM:O00-O9A			
-MONDO:850017	ICD10CM:P00-P96			
-MONDO:850018	ICD10CM:Q00-Q99			
-MONDO:850019	ICD10CM:R00-R99			
-MONDO:850020	ICD10CM:S00-T88			
-MONDO:850021	ICD10CM:U00-U85			
-MONDO:850022	ICD10CM:V00-Y99			
-MONDO:850023	ICD10CM:Z00-Z99			
+mondo_id	mondo_label	xref	xref_source	original_label	definition	parents
+ID	LABEL	A oboInOwl:hasDbXref	>A oboInOwl:source SPLIT=|		A IAO:0000115	SC %
+MONDO:850001	certain infectious and parasitic diseases (a00-b99)	ICD10CM:A00-B99	MONDO:equivalentTo	Certain infectious and parasitic diseases (A00-B99)		
+MONDO:850002	neoplasms (c00-d49)	ICD10CM:C00-D49	MONDO:equivalentTo	Neoplasms (C00-D49)		
+MONDO:850003	diseases of the blood and blood-forming organs and certain disorders involving the immune mechanism (d50-d89)	ICD10CM:D50-D89	MONDO:equivalentTo	Diseases of the blood and blood-forming organs and certain disorders involving the immune mechanism (D50-D89)		
+MONDO:850004	endocrine, nutritional and metabolic diseases (e00-e89)	ICD10CM:E00-E89	MONDO:equivalentTo	Endocrine, nutritional and metabolic diseases (E00-E89)		
+MONDO:850005	mental, behavioral and neurodevelopmental disorders (f01-f99)	ICD10CM:F01-F99	MONDO:equivalentTo	Mental, Behavioral and Neurodevelopmental disorders (F01-F99)		
+MONDO:850006	diseases of the nervous system (g00-g99)	ICD10CM:G00-G99	MONDO:equivalentTo	Diseases of the nervous system (G00-G99)		
+MONDO:850007	diseases of the eye and adnexa (h00-h59)	ICD10CM:H00-H59	MONDO:equivalentTo	Diseases of the eye and adnexa (H00-H59)		
+MONDO:850008	diseases of the ear and mastoid process (h60-h95)	ICD10CM:H60-H95	MONDO:equivalentTo	Diseases of the ear and mastoid process (H60-H95)		
+MONDO:850009	diseases of the circulatory system (i00-i99)	ICD10CM:I00-I99	MONDO:equivalentTo	Diseases of the circulatory system (I00-I99)		
+MONDO:850010	icd-10-cm tabular list of diseases and injuries	ICD10CM:ICD-10-CM	MONDO:equivalentTo	ICD-10-CM TABULAR LIST of DISEASES and INJURIES		
+MONDO:850011	diseases of the respiratory system (j00-j99)	ICD10CM:J00-J99	MONDO:equivalentTo	Diseases of the respiratory system (J00-J99)		
+MONDO:850012	diseases of the digestive system (k00-k95)	ICD10CM:K00-K95	MONDO:equivalentTo	Diseases of the digestive system (K00-K95)		
+MONDO:850013	diseases of the skin and subcutaneous tissue (l00-l99)	ICD10CM:L00-L99	MONDO:equivalentTo	Diseases of the skin and subcutaneous tissue (L00-L99)		
+MONDO:850014	diseases of the musculoskeletal system and connective tissue (m00-m99)	ICD10CM:M00-M99	MONDO:equivalentTo	Diseases of the musculoskeletal system and connective tissue (M00-M99)		
+MONDO:850015	diseases of the genitourinary system (n00-n99)	ICD10CM:N00-N99	MONDO:equivalentTo	Diseases of the genitourinary system (N00-N99)		
+MONDO:850016	pregnancy, childbirth and the puerperium (o00-o9a)	ICD10CM:O00-O9A	MONDO:equivalentTo	Pregnancy, childbirth and the puerperium (O00-O9A)		
+MONDO:850017	certain conditions originating in the perinatal period (p00-p96)	ICD10CM:P00-P96	MONDO:equivalentTo	Certain conditions originating in the perinatal period (P00-P96)		
+MONDO:850018	congenital malformations, deformations and chromosomal abnormalities (q00-q99)	ICD10CM:Q00-Q99	MONDO:equivalentTo	Congenital malformations, deformations and chromosomal abnormalities (Q00-Q99)		
+MONDO:850019	injury, poisoning and certain other consequences of external causes (s00-t88)	ICD10CM:S00-T88	MONDO:equivalentTo	Injury, poisoning and certain other consequences of external causes (S00-T88)		
+MONDO:850020	codes for special purposes (u00-u85)	ICD10CM:U00-U85	MONDO:equivalentTo	Codes for special purposes (U00-U85)