feat: pc owl artifact gen!!

Author: tristan-f-r

datasets/__init__.py

@@ -1,2 +1,3 @@
 /raw
 /intermediate
+/output

datasets/synthetic_data/panther_pathways/.gitignore

@@ -3,7 +3,8 @@ include: "../../../cache/Snakefile"
 rule all:
     input:
         # TODO: pass to script
-        "raw/pathways.txt"   
+        "raw/pathways.txt",
+        "raw/pc-biopax.owl"
 
 produce_fetch_rules({
     "raw/pc-biopax.owl": FetchConfig(["PathwayCommons", "pc-biopax.owl.gz"], uncompress=True),

datasets/synthetic_data/panther_pathways/Snakefile

@@ -1,14 +1,19 @@
 from pathlib import Path
 from paxtools.fetch import fetch
-from ..util.parse_pc_pathways import parse_pc_pathways
+from datasets.synthetic_data.util.parse_pc_pathways import parse_pc_pathways
 
 current_directory = Path(__file__).parent.resolve()
 
 def main():
     pathways_df = parse_pc_pathways(current_directory / 'raw' / 'pathways.txt')
-    print(list(pathways_df[["PATHWAY_URI"]]))
-    # fetch(current_directory / 'raw' / 'pc-biopax.owl', output="pc-biopax-selected.owl")
-    pass
+    print("Fetching pathways... [This may take some time. On the author's desktop machine, it took 15 minutes.]")
+    (current_directory / 'output').mkdir(exist_ok=True)
+    fetch(
+        current_directory / 'raw' / 'pc-biopax.owl',
+        output=(current_directory / 'output' / "pc-biopax-selected.owl"),
+        uris=list(pathways_df["PATHWAY_URI"]),
+        memory=f"{2**(16 - 1)}m" # this is why we don't run this in CI! This is 32gb of memory.
+    )
 
 if __name__ == "__main__":
-    main()
+    main()

datasets/synthetic_data/panther_pathways/fetch_from_owl.py

@@ -31,6 +31,7 @@ package = true
 
 [tool.setuptools.packages.find]
 namespaces = false
+where = ["./"]
 include = ["*"]
 exclude = ["spras*"]