Resolve modules even if packages do not contain inits - allow direct specification of root

pyan/__init__.py

@@ -16,6 +16,7 @@ __version__ = "1.1.2"
 # TODO: fix code duplication with main.py, should have just one implementation.
 def create_callgraph(
     filenames: Union[List[str], str] = "**/*.py",
+    root: str = None,
     function: Union[str, None] = None,
     namespace: Union[str, None] = None,
     format: str = "dot",
@@ -36,6 +37,7 @@ def create_callgraph(
         filenames: glob pattern or list of glob patterns
             to identify filenames to parse (`**` for multiple directories)
             example: **/*.py for all python files
+        root: path to known root directory at which package root sits. Defaults to None, i.e. it will be infered.
         function: if defined, function name to filter for, e.g. "my_module.my_function"
             to only include calls that are related to `my_function`
         namespace: if defined, namespace to filter for, e.g. "my_module", it is highly
@@ -71,7 +73,7 @@ def create_callgraph(
         "annotated": annotated,
     }
 
-    v = CallGraphVisitor(filenames)
+    v = CallGraphVisitor(filenames, root=root)
     if function or namespace:
         if function:
             function_name = function.split(".")[-1]

pyan/analyzer.py

@@ -51,7 +51,7 @@ class CallGraphVisitor(ast.NodeVisitor):
     all files.  This way use information between objects in different files
     can be gathered."""
 
-    def __init__(self, filenames, logger=None):
+    def __init__(self, filenames, root: str = None, logger=None):
         self.logger = logger or logging.getLogger(__name__)
 
         # full module names for all given files
@@ -60,6 +60,7 @@ class CallGraphVisitor(ast.NodeVisitor):
             mod_name = get_module_name(filename)
             self.module_to_filename[mod_name] = filename
         self.filenames = filenames
+        self.root = root
 
         # data gathered from analysis
         self.defines_edges = {}
@@ -103,7 +104,7 @@ class CallGraphVisitor(ast.NodeVisitor):
         with open(filename, "rt", encoding="utf-8") as f:
             content = f.read()
         self.filename = filename
-        self.module_name = get_module_name(filename)
+        self.module_name = get_module_name(filename, root=self.root)
         self.analyze_scopes(content, filename)  # add to the currently known scopes
         self.visit(ast.parse(content, filename))
         self.module_name = None

pyan/anutils.py

@@ -20,27 +20,37 @@ def tail(lst):
         return []
 
 
-def get_module_name(filename):
+def get_module_name(filename, root: str = None):
     """Try to determine the full module name of a source file, by figuring out
-    if its directory looks like a package (i.e. has an __init__.py file)."""
+    if its directory looks like a package (i.e. has an __init__.py file or
+    there is a .py file in it )."""
 
     if os.path.basename(filename) == "__init__.py":
-        return get_module_name(os.path.dirname(filename))
+        # init file means module name is directory name
+        module_path = os.path.dirname(filename)
+    else:
+        # otherwise it is the filename without extension
+        module_path = filename.replace(".py", "")
 
-    init_path = os.path.join(os.path.dirname(filename), "__init__.py")
-    mod_name = os.path.basename(filename).replace(".py", "")
+    # find the module root - walk up the tree and check if it contains .py files - if yes. it is the new root
+    directories = [(module_path, True)]
+    if root is None:
+        while directories[0][0] != os.path.dirname(directories[0][0]):
+            potential_root = os.path.dirname(directories[0][0])
+            is_root = any([f == "__init__.py" for f in os.listdir(potential_root)])
+            directories.insert(0, (potential_root, is_root))
 
-    if not os.path.exists(init_path):
-        return mod_name
+        # keep directories where itself of parent is root
+        while not directories[0][1]:
+            directories.pop(0)
 
-    # blank path means we're looking at __init__.py, in cwd, so its module name is "__init__"
-    if not filename:
-        return "__init__"
+    else:  # root is already known - just walk up until it is matched
+        while directories[0][0] != root:
+            potential_root = os.path.dirname(directories[0][0])
+            directories.insert(0, (potential_root, True))
 
-    if not os.path.dirname(filename):
-        return mod_name
-
-    return get_module_name(os.path.dirname(filename)) + "." + mod_name
+    mod_name = ".".join([os.path.basename(f[0]) for f in directories])
+    return mod_name
 
 
 def format_alias(x):

pyan/main.py

@@ -12,6 +12,7 @@
 from argparse import ArgumentParser
 from glob import glob
 import logging
+import os
 
 from .analyzer import CallGraphVisitor
 from .visgraph import VisualGraph
@@ -149,10 +150,23 @@ def main(cli_args=None):
         help="annotate with module and source line number",
     )
 
+    parser.add_argument(
+        "--root",
+        default=None,
+        dest="root",
+        help="Package root directory. Is inferred by default.",
+    )
+
     known_args, unknown_args = parser.parse_known_args(cli_args)
 
     filenames = [fn2 for fn in unknown_args for fn2 in glob(fn, recursive=True)]
 
+    # determine root
+    if known_args.root is not None:
+        root = os.path.abspath(known_args.root)
+    else:
+        root = None
+
     if len(unknown_args) == 0:
         parser.error("Need one or more filenames to process")
     elif len(filenames) == 0:
@@ -189,7 +203,7 @@ def main(cli_args=None):
         handler = logging.FileHandler(known_args.logname)
         logger.addHandler(handler)
 
-    v = CallGraphVisitor(filenames, logger)
+    v = CallGraphVisitor(filenames, logger, root=root)
 
     if known_args.function or known_args.namespace:
 

pyan/sphinx.py

@@ -74,6 +74,7 @@ class CallgraphDirective(SphinxDirective):
             direction = self.options["direction"]
         dotcode = create_callgraph(
             filenames=f"{base_path}/**/*.py",
+            root=base_path,
             function=func_name,
             namespace=base_name,
             format="dot",

tests/test_analyzer.py

@@ -48,3 +48,17 @@ def test_resolve_use_in_function(callgraph):
     uses = get_in_dict(callgraph.uses_edges, "test_code.submodule2.test_2")
     get_node(uses, "test_code.submodule1.test_func1")
     get_node(uses, "test_code.submodule1.test_func2")
+
+
+def test_resolve_package_without___init__(callgraph):
+    defines = get_in_dict(callgraph.defines_edges, "test_code.subpackage2.submodule_hidden1")
+    get_node(defines, "test_code.subpackage2.submodule_hidden1.test_func1")
+
+
+def test_resolve_package_with_known_root():
+    dirname = os.path.dirname(__file__)
+    filenames = glob(os.path.join(dirname, "test_code/**/*.py"), recursive=True)
+    callgraph = CallGraphVisitor(filenames, logger=logging.getLogger(), root=dirname)
+    dirname_base = os.path.basename(dirname)
+    defines = get_in_dict(callgraph.defines_edges, f"{dirname_base}.test_code.subpackage2.submodule_hidden1")
+    get_node(defines, f"{dirname_base}.test_code.subpackage2.submodule_hidden1.test_func1")