scandeps handles File and Directory objects directly. (#193)

* Scandeps understands File and Directory objects and handles directories
better.

Author: tetron

cwltool/main.py

@@ -472,7 +472,7 @@ def loadref(b, u):
 
     sf = scandeps(
         basedir if basedir else uri, obj, set(("$import", "run")),
-        set(("$include", "$schemas", "path", "location")), loadref)
+        set(("$include", "$schemas")), loadref)
     if sf:
         deps["secondaryFiles"] = sf
 
@@ -491,6 +491,7 @@ def makeRelative(ob):
             else:
                 ob["location"] = os.path.relpath(u, base)
         adjustFileObjs(deps, makeRelative)
+        adjustDirObjs(deps, makeRelative)
 
     stdout.write(json.dumps(deps, indent=4))
 

cwltool/process.py

@@ -1,4 +1,3 @@
-
 import os
 import json
 import copy
@@ -626,14 +625,15 @@ def mergedirs(listing):
         elif e["class"] == "Directory":
             ents[e["basename"]]["listing"].extend(e["listing"])
     for e in ents.itervalues():
-        if e["class"] == "Directory":
+        if e["class"] == "Directory" and "listing" in e:
             e["listing"] = mergedirs(e["listing"])
     r.extend(ents.itervalues())
     return r
 
 def scandeps(base, doc, reffields, urlfields, loadref):
     # type: (Text, Any, Set[Text], Set[Text], Callable[[Text, Text], Any]) -> List[Dict[Text, Text]]
     r = []  # type: List[Dict[Text, Text]]
+    deps = None  # type: Dict[Text, Any]
     if isinstance(doc, dict):
         if "id" in doc:
             if doc["id"].startswith("file://"):
@@ -645,6 +645,16 @@ def scandeps(base, doc, reffields, urlfields, loadref):
                     })
                     base = df
 
+        if doc.get("class") in ("File", "Directory"):
+            u = doc.get("location", doc.get("path"))
+            if u:
+                deps = {
+                    "class": doc["class"],
+                    "location": urlparse.urljoin(base, u)
+                }
+                deps = nestdir(base, deps)
+                r.append(deps)
+
         for k, v in doc.iteritems():
             if k in reffields:
                 for u in aslist(v):
@@ -656,7 +666,7 @@ def scandeps(base, doc, reffields, urlfields, loadref):
                         deps = {
                             "class": "File",
                             "location": subid
-                        }  # type: Dict[Text, Any]
+                        }
                         sf = scandeps(subid, sub, reffields, urlfields, loadref)
                         if sf:
                             deps["secondaryFiles"] = sf