Add new misc. generated code clues

Signed-off-by: Philippe Ombredanne <pombredanne@nexb.com>

Author: pombredanne

src/summarycode/generated.py

@@ -9,7 +9,6 @@
 
 from itertools import islice
 
-
 from commoncode.datautils import Boolean
 from commoncode.text import toascii
 from plugincode.scan import ScanPlugin
@@ -19,7 +18,7 @@
 import typecode.contenttype
 
 """
-Tag files as generated.
+Tag files as generated code based on conspicuous strings.
 """
 
 # Tracing flag
@@ -77,7 +76,7 @@ def generated_scanner(location, **kwargs):
     return dict(is_generated=is_generated)
 
 
-GENERATED_KEYWORDS = tuple(g.lower() for g in (
+GENERATED_KEYWORDS_LOWERED = tuple(g.lower() for g in (
     'generated by',
     'auto-generated',
     'automatically generated',
@@ -160,7 +159,7 @@ def generated_scanner(location, **kwargs):
     'this is an autogenerated file. do not edit this file directly.',
     'this file has been automatically created by',
     'please do not edit this file, all changes will be lost',
-    
+
     # thrift
     'this is an automatically-generated file.',
     'it could get re-generated if the allow_idl_generation flag is on',
@@ -172,22 +171,24 @@ def generated_scanner(location, **kwargs):
 
     # seen in Go files:
     'code generated file. do not edit',
+    'code generated by running "go generate". do not edit.'
 
     # protoc
-    'generated by the protocol buffer compiler.  do not edit!',
-
+    'generated by the protocol buffer compiler. do not edit!',
+    'generated by the protocol buffer compiler.',
 ))
 
 
-max_lines = 150
-
-
-def get_generated_code_hint(location, generated_keywords=GENERATED_KEYWORDS):
+def get_generated_code_hint(
+    location,
+    max_lines=150,
+    generated_keywords=GENERATED_KEYWORDS_LOWERED
+):
     """
     Return a line of extracted text from a file if that file is likely
     generated source code.
 
-    for each of the the first few lines of a source code file
+    for each of the first few lines of a source code file
       if generated keywords are found in the line as lowercase
          yield the line text as a 'potentially_ generated' annotation
     """
@@ -196,9 +197,8 @@ def get_generated_code_hint(location, generated_keywords=GENERATED_KEYWORDS):
         return
     with open(location, 'rb') as filein:
         for line in islice(filein, max_lines):
-            text = toascii(line.strip())
-            textl = text.lower()
-            if any(kw in textl for kw in generated_keywords):
+            text = toascii(line.strip()).lower()
+            if any(kw in text.lower() for kw in generated_keywords):
                 # yield only the first 100 chars..
                 yield text[:100]