Add goal to create general stats

src/ontology/mondo.Makefile

@@ -265,6 +265,56 @@ create-mondo-stats:
 	$(MAKE) create-mondo-stats-summary-file -B
 
 
+#######################################
+# Create General Statistics for Mondo #
+#######################################
+GENERAL_STATISTICS_QUERIES = \
+	$(SPARQLDIR)/reports/COUNT-all_disease_excluding_susceptibility.sparql \
+	$(SPARQLDIR)/reports/COUNT-all_human_diseases.sparql \
+	$(SPARQLDIR)/reports/COUNT-all_non-human_diseases.sparql \
+	$(SPARQLDIR)/reports/COUNT-rare-diseases-classes.sparql \
+	$(SPARQLDIR)/reports/COUNT-human_diseases_infectious.sparql \
+	$(SPARQLDIR)/reports/COUNT-non-human_diseases_infectious.sparql \
+	$(SPARQLDIR)/reports/COUNT-human-genetic-diseases.sparql \
+	$(SPARQLDIR)/reports/COUNT-non-human-genetic-diseases.sparql
+
+TMP_RESULTS_DIR = $(MONDO_STATS_REPORTS_DIR)/tmp
+GEN_STATS_REPORTS_DIR = $(MONDO_STATS_REPORTS_DIR)/mondo-general-stats
+
+OUTPUTS = $(patsubst $(SPARQLDIR)/reports/%.sparql, $(TMP_RESULTS_DIR)/%.tsv, $(GENERAL_STATISTICS_QUERIES))
+
+COMBINED_REPORT = $(GEN_STATS_REPORTS_DIR)/mondo_general_statistics.tsv
+
+create-general-mondo-stats-all: create-general-mondo-stats combine clean-temp
+
+create-general-mondo-stats: $(OUTPUTS)
+
+$(TMP_RESULTS_DIR)/%.tsv: $(SPARQLDIR)/reports/%.sparql mondo.owl
+	mkdir -p $(TMP_RESULTS_DIR) $(GEN_STATS_REPORTS_DIR)
+	@echo "Running query $< ..."
+	$(ROBOT) reason -i mondo.owl query --use-graphs true  -f tsv --query $< $@
+
+# Combine all results into a single report
+combine: create-general-mondo-stats
+	@echo "Combining results into $(COMBINED_REPORT)..."
+	@echo "All Mondo General Statistics created on: $(current_date)" > $(COMBINED_REPORT)
+	cat $(TMP_RESULTS_DIR)/*.tsv >> $(COMBINED_REPORT)
+	@echo "Combined report saved to $(COMBINED_REPORT)"
+
+# Remove temporary result files after combining
+clean-temp:
+	rm -f $(TMP_RESULTS_DIR)/*.tsv
+	@echo "Cleaned up temporary result files."
+
+# Clean everything (temporary + reports)
+clean-stats:
+	rm -f $(TMP_RESULTS_DIR)/*.tsv $(COMBINED_REPORT)
+	@echo "Cleaned all generated files."
+
+all: create-general-mondo-stats-all
+
+
+
 #############################################
 ##### One-time scripts ######################
 #############################################

src/sparql/reports/COUNT-all_disease_excluding_susceptibility.sparql

@@ -0,0 +1,23 @@
+PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
+PREFIX owl: <http://www.w3.org/2002/07/owl#>
+PREFIX obo: <http://purl.obolibrary.org/obo/>
+PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
+
+# Get a count of all non-obsolete disease classes excluding 'disease susceptibility'
+
+SELECT (COUNT(DISTINCT ?class) AS ?countAllDiseaseExcludingSusceptibility)
+WHERE {
+  # Find subclasses of MONDO:0000001 'disease'
+  ?class rdfs:subClassOf* obo:MONDO_0000001 .
+
+  # Exclude subclasses of MONDO:0042489 'disease susceptibility'
+  FILTER NOT EXISTS {
+    ?class rdfs:subClassOf* obo:MONDO_0042489 .
+  }
+
+  FILTER NOT EXISTS {
+    ?class owl:deprecated "true"^^xsd:boolean
+  }
+
+  FILTER (isIRI(?class) && STRSTARTS(str(?class), "http://purl.obolibrary.org/obo/MONDO_"))
+}

src/sparql/reports/COUNT-all_human_diseases.sparql

@@ -0,0 +1,18 @@
+PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
+PREFIX owl: <http://www.w3.org/2002/07/owl#>
+PREFIX obo: <http://purl.obolibrary.org/obo/>
+PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
+
+# Get a count of all non-obsolete 'human disease' classes
+
+SELECT (COUNT(DISTINCT ?class) AS ?countHumanDisease)
+WHERE {
+  # Find subclasses of MONDO:0000001 'disease'
+  ?class rdfs:subClassOf* obo:MONDO_0000001 .
+
+  FILTER NOT EXISTS {
+    ?class owl:deprecated "true"^^xsd:boolean
+  }
+
+  FILTER (isIRI(?class) && STRSTARTS(str(?class), "http://purl.obolibrary.org/obo/MONDO_"))
+}

src/sparql/reports/COUNT-all_non-human_diseases.sparql

@@ -0,0 +1,18 @@
+PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
+PREFIX owl: <http://www.w3.org/2002/07/owl#>
+PREFIX obo: <http://purl.obolibrary.org/obo/>
+PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
+
+# Get a count of all non-obsolete 'non-human animal disease' classes
+
+SELECT (COUNT(DISTINCT ?class) AS ?countNonHumanAnimalDisease)
+WHERE {
+  # Find subclasses of MONDO:0005583 'non-human animal disease'
+  ?class rdfs:subClassOf* obo:MONDO_0005583 .
+
+  FILTER NOT EXISTS {
+    ?class owl:deprecated "true"^^xsd:boolean
+  }
+
+  FILTER (isIRI(?class) && STRSTARTS(str(?class), "http://purl.obolibrary.org/obo/MONDO_"))
+}

src/sparql/reports/COUNT-human-genetic-diseases.sparql

@@ -0,0 +1,22 @@
+PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
+PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
+PREFIX owl: <http://www.w3.org/2002/07/owl#>
+PREFIX obo: <http://purl.obolibrary.org/obo/>
+PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
+
+
+# Get count of human genetic diseases
+
+SELECT (COUNT(DISTINCT ?class) AS ?countHumanGeneticDiseases)
+WHERE {
+  # Find subclasses of MONDO:0003847 'hereditary disease'
+  ?class rdfs:subClassOf* obo:MONDO_0003847 .
+
+  # Exclude deprecated terms
+  FILTER NOT EXISTS {
+    ?class owl:deprecated "true"^^xsd:boolean
+  }
+
+  # Ensure the class is a MONDO term
+  FILTER (isIRI(?class) && STRSTARTS(str(?class), "http://purl.obolibrary.org/obo/MONDO_"))
+}

src/sparql/reports/COUNT-human_diseases_infectious.sparql

@@ -0,0 +1,18 @@
+PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
+PREFIX owl: <http://www.w3.org/2002/07/owl#>
+PREFIX obo: <http://purl.obolibrary.org/obo/>
+PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
+
+# Count the total number of unique non-obsolete Mondo diseases that are an infectious disease
+
+SELECT (COUNT(DISTINCT ?class) AS ?countHumanDiseaseInfectious)
+WHERE {
+  #  'infectious disease' MONDO_0005550
+  ?class rdfs:subClassOf* obo:MONDO_0005550 .
+
+  FILTER NOT EXISTS {
+    ?class owl:deprecated "true"^^xsd:boolean
+  }
+
+  FILTER (isIRI(?class) && STRSTARTS(str(?class), "http://purl.obolibrary.org/obo/MONDO_"))
+}

src/sparql/reports/COUNT-non-human-genetic-diseases.sparql

@@ -0,0 +1,22 @@
+PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
+PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
+PREFIX owl: <http://www.w3.org/2002/07/owl#>
+PREFIX obo: <http://purl.obolibrary.org/obo/>
+PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
+
+
+# Get count of NON-human genetic diseases
+
+SELECT (COUNT(DISTINCT ?class) AS ?countHumanGeneticDiseases)
+WHERE {
+  # Find subclasses of MONDO:1011321 'hereditary disease, non-human animal'
+  ?class rdfs:subClassOf* obo:MONDO_1011321 .
+
+  # Exclude deprecated terms
+  FILTER NOT EXISTS {
+    ?class owl:deprecated "true"^^xsd:boolean
+  }
+
+  # Ensure the class is a MONDO term
+  FILTER (isIRI(?class) && STRSTARTS(str(?class), "http://purl.obolibrary.org/obo/MONDO_"))
+}

src/sparql/reports/COUNT-non-human_diseases_infectious.sparql

@@ -0,0 +1,18 @@
+PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
+PREFIX owl: <http://www.w3.org/2002/07/owl#>
+PREFIX obo: <http://purl.obolibrary.org/obo/>
+PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
+
+# Count the total number of unique non-obsolete Mondo non-human animal diseases that are an infectious disease
+
+SELECT (COUNT(DISTINCT ?class) AS ?countNonHumanDiseaseInfectious)
+WHERE {
+  #  'infectious disease, non-human animal'
+  ?class rdfs:subClassOf* obo:MONDO_0700049 .
+
+  FILTER NOT EXISTS {
+    ?class owl:deprecated "true"^^xsd:boolean
+  }
+
+  FILTER (isIRI(?class) && STRSTARTS(str(?class), "http://purl.obolibrary.org/obo/MONDO_"))
+}

src/sparql/reports/COUNT-rare-diseases-classes.sparql

@@ -0,0 +1,19 @@
+PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
+PREFIX owl: <http://www.w3.org/2002/07/owl#>
+PREFIX obo: <http://purl.obolibrary.org/obo/>
+PREFIX oboInOwl: <http://www.geneontology.org/formats/oboInOwl#>
+PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
+
+# Count of unique non-obsolete Mondo classes that are in the rare disease subset
+
+SELECT (COUNT(DISTINCT ?class) AS ?countRareSubset)
+WHERE {
+  ?class rdfs:subClassOf* obo:MONDO_0000001 ;
+         oboInOwl:inSubset <http://purl.obolibrary.org/obo/mondo#rare> .
+
+  FILTER NOT EXISTS {
+    ?class owl:deprecated "true"^^xsd:boolean
+  }
+
+  FILTER (isIRI(?class) && STRSTARTS(str(?class), "http://purl.obolibrary.org/obo/MONDO_"))
+}