SONARPY-963 Custom stubs should have priority over protobuf generated symbols. Ambigous Symbols of all classes should be disambiguated (#1044)

Author: andrea-guarino-sonarsource

python-frontend/src/main/java/org/sonar/python/types/TypeShed.java

@@ -198,16 +198,16 @@ private static Set<Symbol> searchTypeShedForModule(String moduleName) {
       return new HashSet<>();
     }
     modulesInProgress.add(moduleName);
-    Collection<Symbol> symbolsFromProtobuf = getSymbolsFromProtobufModule(moduleName).values();
-    if (!symbolsFromProtobuf.isEmpty()) {
-      modulesInProgress.remove(moduleName);
-      return new HashSet<>(symbolsFromProtobuf);
-    }
     Set<Symbol> customSymbols = new HashSet<>(getModuleSymbols(moduleName, CUSTOM_THIRD_PARTY, builtinGlobalSymbols).values());
     if (!customSymbols.isEmpty()) {
       modulesInProgress.remove(moduleName);
       return customSymbols;
     }
+    Collection<Symbol> symbolsFromProtobuf = getSymbolsFromProtobufModule(moduleName).values();
+    if (!symbolsFromProtobuf.isEmpty()) {
+      modulesInProgress.remove(moduleName);
+      return new HashSet<>(symbolsFromProtobuf);
+    }
     Set<Symbol> thirdPartySymbols = new HashSet<>(getModuleSymbols(moduleName, THIRD_PARTY_2AND3, builtinGlobalSymbols).values());
     if (thirdPartySymbols.isEmpty()) {
       thirdPartySymbols = commonSymbols(getModuleSymbols(moduleName, THIRD_PARTY_2, builtinGlobalSymbols),
@@ -283,12 +283,28 @@ private static Symbol disambiguateWithLatestPythonSymbol(Set<Symbol> alternative
     return latestPythonSymbol;
   }
 
+
+  /**
+   * Returns stub symbols to be used by SonarSecurity.
+   * Ambiguous symbols that only contain class symbols are disambiguated with latest Python version.
+   */
   public static Collection<Symbol> stubFilesSymbols() {
     Set<Symbol> symbols = new HashSet<>(TypeShed.builtinSymbols().values());
-    typeShedSymbols.values().forEach(symbolsByFqn -> symbols.addAll(symbolsByFqn.values()));
+    for (Map<String, Symbol> symbolsByFqn : typeShedSymbols.values()) {
+      for (Symbol symbol : symbolsByFqn.values()) {
+        symbols.add(isAmbiguousSymbolOfClasses(symbol) ? disambiguateWithLatestPythonSymbol(((AmbiguousSymbol) symbol).alternatives()) : symbol);
+      }
+    }
     return symbols;
   }
 
+  private static boolean isAmbiguousSymbolOfClasses(Symbol symbol) {
+    if (symbol.is(Symbol.Kind.AMBIGUOUS)) {
+      return ((AmbiguousSymbol) symbol).alternatives().stream().allMatch(s -> s.is(Symbol.Kind.CLASS));
+    }
+    return false;
+  }
+
   static class ReturnTypeVisitor extends BaseTreeVisitor {
 
     @Override

python-frontend/src/test/java/org/sonar/python/types/TypeShedTest.java

@@ -488,4 +488,12 @@ public void variables_from_protobuf() throws TextFormat.ParseException {
     assertThat(symbols.values()).extracting(Symbol::kind, Symbol::fullyQualifiedName, Symbol::annotatedTypeName)
       .containsExactlyInAnyOrder(tuple(Kind.OTHER, "mod.foo", "str"), tuple(Kind.OTHER, "mod.bar", null));
   }
+
+  @Test
+  public void stubFilesSymbols_should_not_contain_ambiguous_symbols_of_classes() {
+    // populate typeshed symbols cache by importing socket module
+    symbolsForModule("socket");
+    Collection<Symbol> symbols = TypeShed.stubFilesSymbols();
+    assertThat(symbols).noneMatch(s -> s.is(Kind.AMBIGUOUS) && ((AmbiguousSymbol) s).alternatives().stream().allMatch(a -> a.is(Kind.CLASS)));
+  }
 }