Merge branch 'py_depends' into 'master'

PyUtils: add Python package dependencies to `acmd cmake depends`

See merge request atlas/athena!38812

Tools/PyUtils/CMakeLists.txt

@@ -19,7 +19,7 @@ atlas_install_scripts( bin/acmd.py bin/checkFile.py bin/checkPlugins.py
    bin/diffPoolFiles.py bin/dlldep.py bin/dso-stats.py bin/dump-athfile.py
    bin/dumpAthfilelite.py bin/filter-and-merge-d3pd.py bin/getMetadata.py
    bin/gprof2dot bin/issues bin/magnifyPoolFile.py bin/merge-poolfiles.py
-   bin/pool_extractFileIdentifier.py
+   bin/apydep.py bin/pool_extractFileIdentifier.py
    bin/pool_insertFileToCatalog.py bin/print_auditor_callgraph.py bin/pyroot.py
    bin/vmem-sz.py bin/meta-reader.py bin/meta-diff.py bin/tree-orderer.py
    POST_BUILD_CMD ${ATLAS_FLAKE8} )
@@ -54,3 +54,19 @@ atlas_add_test( RootUtils
 
 atlas_add_test( fprint_test
                 SCRIPT python -m PyUtils.fprint )
+
+
+# Create python package dependencies in release building mode.
+# Used as input for `acmd.py cmake depends`:
+if( NOT "${CMAKE_PROJECT_NAME}" STREQUAL "WorkDir" )
+
+   set( _pydot ${CMAKE_CURRENT_BINARY_DIR}/packages.py.dot )
+   add_custom_command( OUTPUT ${_pydot}
+      COMMAND ${CMAKE_BINARY_DIR}${CMAKE_FILES_DIRECTORY}/atlas_build_run.sh
+      ${CMAKE_CURRENT_SOURCE_DIR}/bin/apydep.py -o ${_pydot} -p ${CMAKE_BINARY_DIR}/${ATLAS_PLATFORM}/packages.txt ${CMAKE_SOURCE_DIR}/../../ )
+
+   add_custom_target( build_pydeps ALL DEPENDS ${_pydot} )
+
+   # Install output if available:
+   install( FILES ${_pydot} DESTINATION . OPTIONAL )
+endif()

Tools/PyUtils/bin/apydep.py

+#!/usr/bin/env python3
+# Copyright (C) 2002-2020 CERN for the benefit of the ATLAS collaboration
+#
+# Created: Oct 2020, Frank Winklmeier
+#
+"""
+Extract Python dependencies between packages and create DOT graph.
+Both `import` and `include` dependencies are considered.
+"""
+
+import ast
+import sys
+import os
+import argparse
+import pygraphviz
+from collections import defaultdict
+
+class DependencyFinder(ast.NodeVisitor):
+   """Walk an AST collecting import/include statements."""
+
+   def __init__(self):
+      self.imports = set()
+      self.includes = set()
+
+   def visit_Import(self, node):
+      """import XYZ"""
+      self.imports.update(alias.name.split('.',1)[0] for alias in node.names)
+
+   def visit_ImportFrom(self, node):
+      """from XYZ import ABC"""
+      if node.level==0:   # ignore relative imports
+         self.imports.add(node.module.split('.',1)[0])
+
+   def visit_Call(self, node):
+      """"include(XYZ/ABC.py)"""
+      if isinstance(node.func, ast.Name) and node.func.id=='include' and node.args:
+         if isinstance(node.args[0], ast.Str):
+            self.includes.add(node.args[0].s.split('/',1)[0])
+
+
+def get_dependencies(filename, print_error=False):
+   """Get all the imports/includes in a file."""
+
+   try:
+      tree = ast.parse(open(filename,'rb').read(), filename=filename)
+   except SyntaxError as e:
+      if print_error:
+         print(e, file=sys.stderr)
+      return DependencyFinder()
+
+   finder = DependencyFinder()
+   finder.visit(tree)
+
+   return finder
+
+
+def walk_tree(path='./', print_error=False, filterFnc=None):
+   """Walk the source tree and extract python dependencies, filtered by FilterFnc"""
+
+   pkg = 'UNKNOWN'
+   deps = defaultdict(lambda : defaultdict(set))
+   for root, dirs, files in os.walk(path):
+      if 'CMakeLists.txt' in files:
+         pkg = os.path.basename(root)
+
+      if (filterFnc and not filterFnc(pkg)):
+         continue
+
+      for f in filter(lambda p : os.path.splitext(p)[1]=='.py', files):
+         d = get_dependencies(os.path.join(root,f), print_error)
+         deps[pkg]['import'].update(d.imports)
+         deps[pkg]['include'].update(d.includes)
+
+   return deps
+
+
+def make_graph(deps, filterFnc=None):
+   """Save the dependencies as dot graph, nodes filtered by filterFnc"""
+
+   graph = pygraphviz.AGraph(name='AthPyGraph', directed=True)
+   for a in deps:
+      for t in ['import','include']:
+         graph.add_edges_from(((a,b) for b in deps[a][t]
+                               if a!=b and (filterFnc is None or (filterFnc(a) and filterFnc(b)))),
+                               label = t)
+   return graph
+
+
+def main():
+   parser = argparse.ArgumentParser(description=__doc__)
+
+   parser.add_argument('path', metavar='DIRECTORY', nargs='?', default='./',
+                       help='root of source tree [%(default)s]')
+
+   parser.add_argument('-o', '--output', metavar='FILE', type=str,
+                       help='output file for DOT graph')
+
+   parser.add_argument('-p', '--packages', metavar='FILE', type=str,
+                       help='path to packages.txt file [from release]')
+
+   parser.add_argument('-a', '--all', action='store_true',
+                       help='include non-athena dependencies')
+
+   parser.add_argument('-v', '--verbose', action='store_true',
+                       help='print parse errors')
+
+   args = parser.parse_args()
+
+   packages = None
+   if not args.all:
+      package_file = args.packages or os.path.join(os.environ['AtlasArea'],'InstallArea',
+                                                   os.environ['BINARY_TAG'],'packages.txt')
+
+      try:
+         with open(package_file) as f:
+            packages = set(line.rstrip().split('/')[-1] for line in f if not line.startswith('#'))
+      except FileNotFoundError:
+         parser.error(f"Cannot read '{package_file}'. Specify via '-p/--packages' or run with '-a/--all'")
+
+   # By default only show athena packages:
+   filterFnc = None if args.all else lambda p : p in packages
+
+   # Walk source tree and create DOT graph:
+   g = make_graph(walk_tree(args.path, args.verbose, filterFnc), filterFnc)
+
+   if args.output:
+      g.write(args.output)
+   else:
+      print(g)
+
+if __name__ == "__main__":
+   sys.exit(main())

Tools/PyUtils/python/scripts/cmake_depends.py

@@ -93,6 +93,16 @@ def subgraph(graph, sources, reverse=False, maxdepth=None, nodegetter=lambda n :
    return g
 
 
+def add_legend(graph):
+   """Add legend to graph"""
+   graph.add_subgraph(name='clusterLegend', label='Legend')
+   l = graph.subgraphs()[-1]
+   for n in 'abcd':
+      l.add_node(n, shape='point', style='invis')
+   l.add_edge('a','b', label='C++', constraint=False)
+   l.add_edge('c','d', label='Python', style='dashed', constraint=False)
+
+
 class AthGraph:
    """Class to hold dependency information for release"""
 
@@ -104,13 +114,13 @@ class AthGraph:
 
       # Build dictionary for node types:
       legend = self.graph.get_subgraph('clusterLegend')
-      self.types = { n.attr['label'] : n.attr['shape'] for n in legend.iternodes() }
+      self.types = { n.attr['label'] : n.attr['shape'] for n in legend.nodes_iter() }
 
       # Build dictionary for node names:
-      self.node = { n.attr['label'] : n.get_name() for n in self.graph.iternodes() }
+      self.node = { n.attr['label'] : n.get_name() for n in self.graph.nodes_iter() }
 
       # Extract package dependencies:
-      for e in self.graph.iteredges():
+      for e in self.graph.edges_iter():
          p = e[0].attr['label']
          # Decorate target with package name:
          if p.startswith('Package_'):
@@ -119,7 +129,7 @@ class AthGraph:
 
       # Assign "package" names to externals if possible:
       external_nodes = filter(lambda n : 'package' not in n.attr.keys(),
-                              self.graph.iternodes())
+                              self.graph.nodes_iter())
       for n in external_nodes:
          name = externals_name(n.attr['label'])
          n.attr['package'] = name.split('::')[0]
@@ -162,6 +172,9 @@ class AthGraph:
                   type=int, default=1, const=None,
                   help='recursively resolve dependencies up to DEPTH (default: unlimited)')
 
+@acmdlib.argument('--py', action='store_true',
+                  help='add Python dependencies')
+
 @acmdlib.argument('--regex', action='store_true',
                   help='treat NAME as regular expression')
 
@@ -171,8 +184,14 @@ class AthGraph:
 @acmdlib.argument('-d', '--dot', action='store_true',
                   help='print DOT graph')
 
+@acmdlib.argument('--noLegend', action='store_true',
+                  help='do not add legend to graph')
+
+
 # Debugging/expert options:
 @acmdlib.argument('--cmakedot', help=argparse.SUPPRESS)
+@acmdlib.argument('--pydot', help=argparse.SUPPRESS)
+
 
 def main(args):
    """Inspect cmake build dependencies"""
@@ -185,6 +204,18 @@ def main(args):
       except KeyError:
          main.parser.error("Cannot find 'packages.dot'. Setup a release or use --cmakedot.")
 
+   # Find packages.py.dot:
+   pygraph = None
+   if args.py:
+      if args.target:
+         main.parser.error("Python dependencies not possible in target mode.")
+
+      args.pydot = args.pydot or args.cmakedot.replace('.dot','.py.dot')
+      try:
+         pygraph = pygraphviz.AGraph(args.pydot)
+      except Exception:
+         main.parser.error(f"Cannot read '{args.pydot}'. Setup a release or use --pydot.")
+
    # Read packages.txt if needed:
    package_paths = {}
    if args.long:
@@ -195,8 +226,9 @@ def main(args):
          main.parser.error("Cannot read 'packages.txt'. Setup a release or run without -l/--long.")
 
    # In package mode we have one extra level due to the Package_ target:
+   depth = args.recursive
    if not args.target and not args.clients and args.recursive is not None:
-      args.recursive += 1
+      depth += 1
 
    # Read dependencies:
    d = AthGraph(args.cmakedot, package_paths)
@@ -208,14 +240,14 @@ def main(args):
          a = 'label' if args.target else 'package'
          return node.attr[a]
 
-   graph = pygraphviz.AGraph(name='AthGraph', directed=True)
+   graph = pygraphviz.AGraph(name='AthGraph', directed=True, strict=False)
    for p in args.names:
       target = p.split('/')[-1]  # in case of full package path
 
       # With regex, find all matching targets:
       if args.regex:
          r = re.compile(target)
-         targets = [getnode(n) for n in d.graph.iternodes() if r.match(n.attr['label'])]
+         targets = [getnode(n) for n in d.graph.nodes_iter() if r.match(n.attr['label'])]
          targets = list(filter(lambda t : t is not None, targets))
       else:
          targets = [target]
@@ -242,12 +274,31 @@ def main(args):
 
       # Extract the dependency subgraph:
       g = subgraph(d.graph, sources, reverse=args.clients,
-                   maxdepth=args.recursive, nodegetter=getnode)
+                   maxdepth=depth, nodegetter=getnode)
 
       graph.add_subgraph(name=target)
       graph.get_subgraph(target).add_edges_from(g.edges())
 
-   # Print result:
+      # Add python dependencies:
+      if args.py:
+         # Here the nodes are the actual package names:
+         pysources = [pygraph.get_node(t) for t in targets if pygraph.has_node(t)]
+         g = subgraph(pygraph, pysources, reverse=args.clients,
+                      maxdepth=args.recursive, nodegetter=lambda n : n.name)
+
+         graph.get_subgraph(target).add_edges_from(g.edges(), style='dashed')
+
+         # Change style of nodes that have only Python dependencies:
+         g = graph.get_subgraph(target)
+         for n in g.nodes_iter():
+            if all(e.attr['style']=='dashed' for e in g.edges_iter(n)):
+               n.attr['style'] = 'dashed'
+
+
+   # Output final graph:
+   if not args.noLegend:
+      add_legend(graph)
+
    if args.dot:
       print(graph)
    else: