more type improvements (#72)

Author: mr-c

cwltool/__main__.py

@@ -1,4 +1,4 @@
-import main
+from . import main
 import sys
 
 sys.exit(main.main())

cwltool/builder.py

@@ -33,20 +33,20 @@ class Builder(object):
 
     def __init__(self):  # type: () -> None
         self.names = None  # type: avro.schema.Names
-        self.schemaDefs = None  # type: Dict[str,Dict[str,str]]
-        self.files = None  # type: List[str]
+        self.schemaDefs = None  # type: Dict[str,Dict[unicode, Any]]
+        self.files = None  # type: List[Dict[str, str]]
         self.fs_access = None  # type: StdFsAccess
-        self.job = None  # type: Dict[str,str]
+        self.job = None  # type: Dict[str, Any]
         self.requirements = None  # type: List[Dict[str,Any]]
         self.outdir = None  # type: str
         self.tmpdir = None  # type: str
-        self.resources = None  # type: Dict[str,str]
-        self.bindings = []  # type: List[Dict[str,str]]
+        self.resources = None  # type: Dict[str, Union[int, str]]
+        self.bindings = []  # type: List[Dict[str, Any]]
         self.timeout = None  # type: int
         self.pathmapper = None # type: PathMapper
 
     def bind_input(self, schema, datum, lead_pos=[], tail_pos=[]):
-        # type: (Dict[str,Any], Any, List[int], List[int]) -> List[Dict[str,str]]
+        # type: (Dict[unicode, Any], Any, List[int], List[int]) -> List[Dict[str, Any]]
         bindings = []  # type: List[Dict[str,str]]
         binding = None  # type: Dict[str,Any]
         if "inputBinding" in schema and isinstance(schema["inputBinding"], dict):
@@ -64,7 +64,7 @@ def bind_input(self, schema, datum, lead_pos=[], tail_pos=[]):
         # Handle union types
         if isinstance(schema["type"], list):
             for t in schema["type"]:
-                if isinstance(t, basestring) and self.names.has_name(t, ""):
+                if isinstance(t, (str, unicode)) and self.names.has_name(t, ""):
                     avsc = self.names.get_name(t, "")
                 elif isinstance(t, dict) and "name" in t and self.names.has_name(t["name"], ""):
                     avsc = self.names.get_name(t["name"], "")
@@ -148,7 +148,7 @@ def _capture_files(f):
 
         return bindings
 
-    def tostr(self, value):  # type(Any) -> str
+    def tostr(self, value):  # type: (Any) -> str
         if isinstance(value, dict) and value.get("class") == "File":
             if "path" not in value:
                 raise WorkflowException(u"File object must have \"path\": %s" % (value))

cwltool/cwlrdf.py

@@ -1,16 +1,19 @@
 import json
 import urlparse
+from schema_salad.ref_resolver import Loader
 from rdflib import Graph, plugin, URIRef
 from rdflib.serializer import Serializer
 from typing import Any, Union, Dict, IO
 
 def makerdf(workflow, wf, ctx):
-    # type: (str, Dict[str,Any], Dict[str,Union[str, Dict[str,str]]]) -> Graph
+    # type: (str, Dict[str,Any], Loader.ContextType) -> Graph
     prefixes = {}
     for k,v in ctx.iteritems():
         if isinstance(v, dict):
-            v = v["@id"]
-        doc_url, frg = urlparse.urldefrag(v)
+            url = v["@id"]
+        else:
+            url = v
+        doc_url, frg = urlparse.urldefrag(url)
         if "/" in frg:
             p, _ = frg.split("/")
             prefixes[p] = u"%s#%s/" % (doc_url, p)
@@ -22,13 +25,13 @@ def makerdf(workflow, wf, ctx):
     for s,p,o in g.triples((None, URIRef("@id"), None)):
         g.remove((s, p, o))
 
-    for k,v in prefixes.iteritems():
-        g.namespace_manager.bind(k, v)
+    for k2,v2 in prefixes.iteritems():
+        g.namespace_manager.bind(k2, v2)
 
     return g
 
 def printrdf(workflow, wf, ctx, sr, stdout):
-    # type: (str, Dict[str,Any], Dict[str,Union[str, Dict[str,str]]], str, IO[Any]) -> None
+    # type: (str, Dict[str,Any], Loader.ContextType, str, IO[Any]) -> None
     stdout.write(makerdf(workflow, wf, ctx).serialize(format=sr))
 
 def lastpart(uri):  # type: (Any) -> str
@@ -158,7 +161,7 @@ def dot_without_parameters(g, stdout):  # type: (Graph, IO[Any]) -> None
            }""")
 
     for src, sink, srcrun, sinkrun in qres:
-        attr = ""
+        attr = u""
         if srcrun in clusternode:
             attr += u'ltail="%s"' % dotname[srcrun]
             src = clusternode[srcrun]
@@ -169,7 +172,7 @@ def dot_without_parameters(g, stdout):  # type: (Graph, IO[Any]) -> None
 
 
 def printdot(workflow, wf, ctx, stdout, include_parameters=False):
-    # type: (str, Dict[str,Any], Dict[str,Union[str, Dict[str,str]]], Any, bool) -> None
+    # type: (str, Dict[str,Any], Loader.ContextType, Any, bool) -> None
     g = makerdf(workflow, wf, ctx)
 
     stdout.write("digraph {")

cwltool/cwltest.py

@@ -165,9 +165,12 @@ def main():  # type: () -> int
         tests = []
         for t in alltests:
             loader = schema_salad.ref_resolver.Loader({"id": "@id"})
-            cwl, _ = loader.resolve_ref(t["tool"])
-            if cwl["class"] == "CommandLineTool":
-                tests.append(t)
+            cwl = loader.resolve_ref(t["tool"])[0]
+            if isinstance(cwl, dict):
+                if cwl["class"] == "CommandLineTool":
+                    tests.append(t)
+            else:
+                raise Exception("Unexpected code path.")
 
     if args.l:
         for i, t in enumerate(tests):

cwltool/draft2tool.py

@@ -100,7 +100,7 @@ def __init__(self, toolpath_object, **kwargs):
         # type: (Dict[str,Any], **Any) -> None
         super(CommandLineTool, self).__init__(toolpath_object, **kwargs)
 
-    def makeJobRunner(self):
+    def makeJobRunner(self):  # type: () -> CommandLineJob
         return CommandLineJob()
 
     def makePathMapper(self, reffiles, input_basedir, **kwargs):
@@ -267,7 +267,7 @@ def collect_output_ports(self, ports, builder, outdir):
             raise WorkflowException("Error validating output record, " + str(e) + "\n in " + json.dumps(ret, indent=4))
 
     def collect_output(self, schema, builder, outdir):
-        # type: (Dict[str,Any],Builder,str) -> Union[Dict[str,Any],List[Union[Dict[str,Any],str]]]
+        # type: (Dict[str,Any], Builder, str) -> Union[Dict[str, Any], List[Union[Dict[str, Any], str]]]
         r = []  # type: List[Any]
         if "outputBinding" in schema:
             binding = schema["outputBinding"]
@@ -372,9 +372,9 @@ def collect_output(self, schema, builder, outdir):
 
         if (not r and isinstance(schema["type"], dict) and
                 schema["type"]["type"] == "record"):
-            out = {} # type: Dict[str, Any]
+            out = {}
             for f in schema["type"]["fields"]:
-                out[shortname(f["name"])] = self.collect_output(
+                out[shortname(f["name"])] = self.collect_output(  # type: ignore
                         f, builder, outdir)
             return out
         return r

cwltool/expression.py

@@ -15,14 +15,14 @@
 _logger = logging.getLogger("cwltool")
 
 def jshead(engineConfig, rootvars):
-    # type: (List[str],Dict[str,str]) -> str
-    return "\n".join(engineConfig + [u"var %s = %s;" % (k, json.dumps(v, indent=4)) for k, v in rootvars.items()])
+    # type: (List[unicode],Dict[str,str]) -> unicode
+    return u"\n".join(engineConfig + [u"var %s = %s;" % (k, json.dumps(v, indent=4)) for k, v in rootvars.items()])
 
 def exeval(ex, jobinput, requirements, outdir, tmpdir, context, pull_image):
     # type: (Dict[str,Any], Dict[str,str], List[Dict[str, Any]], str, str, Any, bool) -> sandboxjs.JSON
 
     if ex["engine"] == "https://w3id.org/cwl/cwl#JavascriptEngine":
-        engineConfig = []  # type: List[str]
+        engineConfig = []  # type: List[unicode]
         for r in reversed(requirements):
             if r["class"] == "ExpressionEngineRequirement" and r["id"] == "https://w3id.org/cwl/cwl#JavascriptEngine":
                 engineConfig = r.get("engineConfig", [])
@@ -126,7 +126,7 @@ def param_interpolate(ex, obj, strip=True):
 
 def do_eval(ex, jobinput, requirements, outdir, tmpdir, resources,
             context=None, pull_image=True, timeout=None):
-    # type: (Any, Dict[str,str], List[Dict[str,Any]], str, str, Dict[str,str], Any, bool, int) -> Any
+    # type: (Any, Dict[str,str], List[Dict[str,Any]], str, str, Dict[str, Union[int, str]], Any, bool, int) -> Any
 
     runtime = resources.copy()
     runtime["tmpdir"] = tmpdir

cwltool/job.py

@@ -48,10 +48,10 @@ def __init__(self):  # type: () -> None
         self.successCodes = None  # type: Iterable[int]
         self.temporaryFailCodes = None  # type: Iterable[int]
         self.permanentFailCodes = None  # type: Iterable[int]
-        self.requirements = None  # type: Dict[str,str]
+        self.requirements = None  # type: List[Dict[str, str]]
         self.hints = None  # type: Dict[str,str]
-        self.name = None  # type: str
-        self.command_line = None  # type: List[str]
+        self.name = None  # type: unicode
+        self.command_line = None  # type: List[unicode]
         self.pathmapper = None  # type: PathMapper
         self.collect_outputs = None  # type: Union[Callable[[Any], Any],functools.partial[Any]]
         self.output_callback = None  # type: Callable[[Any, Any], Any]
@@ -69,7 +69,7 @@ def run(self, dry_run=False, pull_image=True, rm_container=True,
         #with open(os.path.join(outdir, "cwl.input.json"), "w") as fp:
         #    json.dump(self.joborder, fp)
 
-        runtime = []  # type: List[str]
+        runtime = []  # type: List[unicode]
         env = {"TMPDIR": self.tmpdir}  # type: Mapping[str,str]
 
         (docker_req, docker_is_req) = get_feature(self, "DockerRequirement")

cwltool/main.py

@@ -22,7 +22,7 @@
 from .process import shortname, Process
 import rdflib
 from .utils import aslist
-from typing import Union, Any, cast, Callable, Tuple, IO
+from typing import Union, Any, cast, Callable, Dict, Tuple, IO
 
 _logger = logging.getLogger("cwltool")
 
@@ -294,15 +294,15 @@ def load_tool(argsworkflow, updateonly, strict, makeTool, debug,
               rdf_serializer=None,
               stdout=sys.stdout,
               urifrag=None):
-    # type: (Union[str,unicode,dict[str,Any]], bool, bool, Callable[...,Process], bool, bool, bool, bool, bool, bool, Any, Any, Any) -> Any
+    # type: (Union[str,unicode,dict[unicode,Any]], bool, bool, Callable[...,Process], bool, bool, bool, bool, bool, bool, Any, Any, Any) -> Any
     (document_loader, avsc_names, schema_metadata) = process.get_schema()
 
     if isinstance(avsc_names, Exception):
         raise avsc_names
 
     jobobj = None
     uri = None  # type: str
-    workflowobj = None  # type: Dict[str, Any]
+    workflowobj = None  # type: Dict[unicode, Any]
     if isinstance(argsworkflow, (basestring)):
         split = urlparse.urlsplit(cast(str, argsworkflow))
         if split.scheme:
@@ -343,9 +343,11 @@ def load_tool(argsworkflow, updateonly, strict, makeTool, debug,
         return 0
 
     try:
-        processobj, metadata = schema_salad.schema.load_and_validate(document_loader, avsc_names, workflowobj, strict)
+        processobj, metadata = schema_salad.schema.load_and_validate(
+                document_loader, avsc_names, workflowobj, strict)
     except (schema_salad.validate.ValidationException, RuntimeError) as e:
-        _logger.error(u"Tool definition failed validation:\n%s", e, exc_info=(e if debug else False))
+        _logger.error(u"Tool definition failed validation:\n%s", e,
+                exc_info=(e if debug else False))
         return 1
 
     if print_pre:
@@ -401,7 +403,10 @@ def load_job_order(args, t, parser, stdin, print_input_deps=False, relative_deps
     if args.conformance_test:
         loader = Loader({})
     else:
-        jobloaderctx = {"path": {"@type": "@id"}, "format": {"@type": "@id"}, "id": "@id"}
+        jobloaderctx = {
+                "path": {"@type": "@id"},
+                "format": {"@type": "@id"},
+                "id": "@id"}
         jobloaderctx.update(t.metadata.get("$namespaces", {}))
         loader = Loader(jobloaderctx)
 
@@ -478,7 +483,7 @@ def load_job_order(args, t, parser, stdin, print_input_deps=False, relative_deps
 
 
 def printdeps(obj, document_loader, stdout, relative_deps, basedir=None):
-    # type: (Dict[str,Any], Loader, IO[Any], bool, str) -> None
+    # type: (Dict[unicode, Any], Loader, IO[Any], bool, str) -> None
     deps = {"class": "File",
             "path": obj.get("id", "#")}
 
@@ -507,7 +512,7 @@ def makeRelative(u):
     stdout.write(json.dumps(deps, indent=4))
 
 def versionstring():
-    # type: () -> str
+    # type: () -> unicode
     pkg = pkg_resources.require("cwltool")
     if pkg:
         return u"%s %s" % (sys.argv[0], pkg[0].version)
@@ -524,7 +529,7 @@ def main(argsl=None,
          stdout=sys.stdout,
          stderr=sys.stderr,
          versionfunc=versionstring):
-    # type: (List[str],Callable[...,Union[str,Dict[str,str]]],Callable[...,Process],Callable[[Dict[str,int]],Dict[str,int]],argparse.ArgumentParser,IO[Any],IO[Any],IO[Any],Callable[[],str]) -> int
+    # type: (List[str],Callable[...,Union[str,Dict[str,str]]],Callable[...,Process],Callable[[Dict[str,int]],Dict[str,int]],argparse.ArgumentParser,IO[Any],IO[Any],IO[Any],Callable[[],unicode]) -> int
 
     _logger.removeHandler(defaultStreamHandler)
     _logger.addHandler(logging.StreamHandler(stderr))

cwltool/pathmapper.py

@@ -20,13 +20,12 @@ class PathMapper(object):
     """Mapping of files from relative path provided in the file to a tuple of
     (absolute local path, absolute container path)"""
 
-    def __new__(cls, referenced_files, basedir, *args, **kwargs):
-        # type: (Set[str], str) -> Any
-        instance = super(PathMapper,cls).__new__(cls)
-        instance._pathmap = {}  # type: Dict[str, Tuple[str, str]]
-        return instance
-
     def __init__(self, referenced_files, basedir):
+        # type: (Set[str], str) -> None
+        self._pathmap = {}  # type: Dict[str, Tuple[str, str]]
+        self.setup(referenced_files, basedir)
+
+    def setup(self, referenced_files, basedir):
         # type: (Set[str], str) -> None
         for src in referenced_files:
             ab = abspath(src, basedir)
@@ -52,30 +51,28 @@ def reversemap(self, target):  # type: (str) -> Tuple[str, str]
 
 class DockerPathMapper(PathMapper):
 
-    def __new__(cls, referenced_files, basedir):
-        # type: (Set[str], str) -> None
-        instance = super(DockerPathMapper,cls).__new__(cls, referenced_files, basedir)
-        instance.dirs = {}  # type: Dict[str, Union[bool, str]]
-        return instance
-
     def __init__(self, referenced_files, basedir):
         # type: (Set[str], str) -> None
+        self.dirs = {}  # type: Dict[str, Union[bool, str]]
+        super(DockerPathMapper, self).__init__(referenced_files, basedir)
+
+    def setup(self, referenced_files, basedir):
         for src in referenced_files:
             ab = abspath(src, basedir)
-            dir, fn = os.path.split(ab)
+            dirn, fn = os.path.split(ab)
 
             subdir = False
             for d in self.dirs:
-                if dir.startswith(d):
+                if dirn.startswith(d):
                   subdir = True
                   break
 
             if not subdir:
                 for d in list(self.dirs):
-                    if d.startswith(dir):
-                        # 'dir' is a parent of 'd'
+                    if d.startswith(dirn):
+                        # 'dirn' is a parent of 'd'
                         del self.dirs[d]
-                self.dirs[dir] = True
+                self.dirs[dirn] = True
 
         prefix = "job" + str(random.randint(1, 1000000000)) + "_"
 
@@ -85,7 +82,8 @@ def __init__(self, referenced_files, basedir):
             i = 1
             while name in names:
                 i += 1
-                name = os.path.join("/var/lib/cwl", prefix + os.path.basename(d) + str(i))
+                name = os.path.join("/var/lib/cwl",
+                        prefix + os.path.basename(d) + str(i))
             names.add(name)
             self.dirs[d] = name
 
@@ -96,9 +94,11 @@ def __init__(self, referenced_files, basedir):
             st = os.lstat(deref)
             while stat.S_ISLNK(st.st_mode):
                 rl = os.readlink(deref)
-                deref = rl if os.path.isabs(rl) else os.path.join(os.path.dirname(deref), rl)
+                deref = rl if os.path.isabs(rl) else os.path.join(
+                        os.path.dirname(deref), rl)
                 st = os.lstat(deref)
 
             for d in self.dirs:
                 if ab.startswith(d):
-                    self._pathmap[src] = (deref, os.path.join(self.dirs[d], ab[len(d)+1:]))
+                    self._pathmap[src] = (deref, os.path.join(
+                        self.dirs[d], ab[len(d)+1:]))