SLOC and comment stats

sloc.py

import os, sys

c_like = {"single": "//", "multi_start": ("/*",), "multi_end": ("*/",),}

langs = {
    "py": {"single": "#", "multi_start": ("'''", '"""'),
        "multi_end": ("'''", '"""'),},
    "html": {"single": "//", "multi_start": ("<!--", "/*"),
        "multi_end": ("-->", "*/"),},
    "js": c_like,
    "scala": c_like,
    "css": {"single": None, "multi_start": ("/*",), "multi_end": ("*/",),},
}

for lang in langs:
    langs[lang].update({"sloc": 0, "comments": 0, "files": 0, })
    
in_comment = False
root = "."
if len(sys.argv) > 1:
    root = sys.argv[1]
if not os.path.isdir(root):
    print
    print "%s is not a directory" % root
    print
    sys.exit()

for (dirpath, dirnames, filenames) in os.walk(root):
    for filename in filenames:
        lang = langs.get(filename.split(".")[-1])
        if lang:
            filename = os.path.join(dirpath, filename)
            print "Reading %s" % filename
            with open(filename, "r") as f:
                lang["files"] += 1
                for line in f:
                    line = line.strip()
                    if line:
                        if in_comment:
                            for end in lang["multi_end"]:
                                if line.endswith(end):
                                    # End of multi-line comment.
                                    in_comment = False
                                    if line.replace(end, "").strip():
                                        lang["comments"] += 1
                                    break
                            else:
                                # Inside multi-line comment.
                                lang["comments"] += 1
                        elif (lang["single"] and 
                            line.startswith(lang["single"])):
                            # Single line comment.
                            lang["comments"] += 1
                        else:
                            for start in lang["multi_start"]:
                                if line.startswith(start):
                                    # Start of multi-line comment.
                                    in_comment = True
                                    if line.replace(start, "").strip():
                                        lang["comments"] += 1
                                    break
                            else:
                                # Line of code.
                                lang["sloc"] += 1

def print_result(title, data):
    print
    print "=" * 20
    print title.upper()
    print "=" * 20
    print "Files: %s" % data["files"]
    print "SLOC: %s" % data["sloc"]
    print "Lines of comments: %s" % data["comments"]
    print "Total non-blank lines: %s" % (data["sloc"] + data["comments"])
    if data["comments"] > 0:
        ratio = "%s:1" % (data["sloc"] / data["comments"])
    else:
        ratio = 0
    print "SLOC per lines of comments: %s" % ratio

def total(count):
    return sum([lang[count] for lang in langs.values()])
    
for name, lang in langs.items():
    print_result(name, lang)

counts = ("files", "sloc", "comments")
total = lambda count: sum([lang[count] for lang in langs.values()])
totals = dict([(count, total(count)) for count in counts])
print_result("Total", totals)
print