filter browseable files/dirs to h5ad, 10x h5, Xenium, loom, Seurat, T…

…ileDB, or zarr formats

cirrocumulus/api.py

@@ -168,12 +168,42 @@ def handle_server():
         other=["Gallus gallus", "Macaca fascicularis", "Macaca mulatta", "Rattus norvegicus"],
     )
 
-    # browse server-side data files to allow user to select the desired dataset instead of necessarily typing the URL
+    # browse server-side data files to allow user to select the desired dataset instead of necessarily typing the URL:
     d["server_files"] = []
     if CIRRO_SERVER_DATA_DIR in os.environ and os.environ[CIRRO_SERVER_DATA_DIR]:
+        VALID_EXTENSIONS = [
+            ".h5ad",
+            ".h5",
+            ".zip",
+            ".tar",
+            ".tar.gz",
+            ".loom",
+            ".h5seurat",
+            ".rds",
+            ".zarr",
+        ]
         for root, dirs, files in os.walk(os.environ[CIRRO_SERVER_DATA_DIR]):
             for file in dirs + files:
-                d["server_files"] += [os.path.join(root, file)]
+                if any([file.endswith(ext) for ext in VALID_EXTENSIONS]):
+                    d["server_files"] += [os.path.join(root, file)]
+                    # if this is a directory, e.g. .zarr "files", prevent further recursion:
+                    if file in dirs:
+                        dirs.remove(file)
+                elif (
+                    file in dirs
+                ):  # check if `file` is a MEX formatted directory (we must look at the subfiles)
+                    # https://www.10xgenomics.com/support/software/xenium-panel-designer/latest/tutorials/create-single-cell-reference
+                    count_tsvgz = 0  # MEX directoris have two subfiles with .tsv.gz extension
+                    count_mtxgz = 0  # MEX directories also have one subfile with .mtx.gz extension
+                    for subfile in os.listdir(os.path.join(root, file)):
+                        if subfile.endswith(".tsv.gz"):
+                            count_tsvgz += 1
+                        elif subfile.endswith(".mtx.gz"):
+                            count_mtxgz += 1
+                        if count_tsvgz == 2 and count_mtxgz == 1:
+                            d["server_files"] += [os.path.join(root, file)]
+                            dirs.remove(file)
+                            break
 
     # from https://www.ebi.ac.uk/ols/ontologies/efo/terms?iri=http%3A%2F%2Fwww.ebi.ac.uk%2Fefo%2FEFO_0010183&viewMode=All&siblings=false
     d["library"] = load_json(CIRRO_LIBRARY) or [