# Licensed under the Apache License: http://www.apache.org/licenses/LICENSE-2.0
# For details: https://github.com/nedbat/django_coverage_plugin/blob/master/NOTICE.txt

"""The Django template coverage plugin."""

from __future__ import print_function

import os.path
import re

import coverage.plugin
import django
import django.template
from django.template.base import Lexer, NodeList, Template, TextNode
from django.template.defaulttags import VerbatimNode
from django.templatetags.i18n import BlockTranslateNode
from six.moves import range

try:
    from django.template.base import TokenType

    def _token_name(token_type):
        token_type.name.capitalize()

except ImportError:
    # Django <2.1 uses separate constants for token types
    from django.template.base import (
        TOKEN_BLOCK, TOKEN_MAPPING, TOKEN_TEXT, TOKEN_VAR
    )

    class TokenType:
        TEXT = TOKEN_TEXT
        VAR = TOKEN_VAR
        BLOCK = TOKEN_BLOCK

    def _token_name(token_type):
        return TOKEN_MAPPING[token_type]


class DjangoTemplatePluginException(Exception):
    """Used for any errors from the plugin itself."""
    pass


# For debugging the plugin itself.
SHOW_PARSING = False
SHOW_TRACING = False


def check_debug():
    """Check that Django's template debugging is enabled.

    Django's built-in "template debugging" records information the plugin needs
    to do its work.  Check that the setting is correct, and raise an exception
    if it is not.

    Returns True if the debug check was performed, False otherwise
    """
    from django.conf import settings

    if not settings.configured:
        return False

    # I _think_ this check is all that's needed and the 3 "hasattr" checks
    # below can be removed, but it's not clear how to verify that
    from django.apps import apps
    if not apps.ready:
        return False

    # django.template.backends.django gets loaded lazily, so return false
    # until they've been loaded
    if not hasattr(django.template, "backends"):
        return False
    if not hasattr(django.template.backends, "django"):
        return False
    if not hasattr(django.template.backends.django, "DjangoTemplates"):
        raise DjangoTemplatePluginException("Can't use non-Django templates.")

    for engine in django.template.engines.all():
        if not isinstance(engine, django.template.backends.django.DjangoTemplates):
            raise DjangoTemplatePluginException(
                "Can't use non-Django templates."
            )
        if not engine.engine.debug:
            raise DjangoTemplatePluginException(
                "Template debugging must be enabled in settings."
            )

    return True


if django.VERSION < (1, 8):
    raise RuntimeError("Django Coverage Plugin requires Django 1.8 or higher")


if django.VERSION >= (1, 9):
    # Since we are grabbing at internal details, we have to adapt as they
    # change over versions.
    def filename_for_frame(frame):
        try:
            return frame.f_locals["self"].origin.name
        except (KeyError, AttributeError):
            return None

    def position_for_node(node):
        try:
            return node.token.position
        except AttributeError:
            return None

    def position_for_token(token):
        return token.position
else:
    def filename_for_frame(frame):
        try:
            return frame.f_locals["self"].source[0].name
        except (KeyError, AttributeError, IndexError):
            return None

    def position_for_node(node):
        return node.source[1]

    def position_for_token(token):
        return token.source[1]


def read_template_source(filename):
    """Read the source of a Django template, returning the Unicode text."""
    # Import this late to be sure we don't trigger settings machinery too
    # early.
    from django.conf import settings

    if not settings.configured:
        settings.configure()

    with open(filename, "rb") as f:
        # The FILE_CHARSET setting will be removed in 3.1:
        # https://docs.djangoproject.com/en/3.0/ref/settings/#file-charset
        if django.VERSION >= (3, 1):
            charset = 'utf-8'
        else:
            charset = settings.FILE_CHARSET
        text = f.read().decode(charset)

    return text


class DjangoTemplatePlugin(
    coverage.plugin.CoveragePlugin,
    coverage.plugin.FileTracer,
):

    def __init__(self):
        self.debug_checked = False

        self.django_template_dir = os.path.realpath(
            os.path.dirname(django.template.__file__)
        )

        self.source_map = {}

    # --- CoveragePlugin methods

    def sys_info(self):
        return [
            ("django_template_dir", self.django_template_dir),
            ("environment", sorted(
                ("%s = %s" % (k, v))
                for k, v in os.environ.items()
                if "DJANGO" in k
            )),
        ]

    def file_tracer(self, filename):
        if filename.startswith(self.django_template_dir):
            if not self.debug_checked:
                # Keep calling check_debug until it returns True, which it
                # will only do after settings have been configured
                self.debug_checked = check_debug()

            return self
        return None

    def file_reporter(self, filename):
        return FileReporter(filename)

    def find_executable_files(self, src_dir):
        for (dirpath, dirnames, filenames) in os.walk(src_dir):
            for filename in filenames:
                # We're only interested in files that look like reasonable HTML
                # files: Must end with .htm or .html, and must not have certain
                # funny characters that probably mean they are editor junk.
                if re.match(r"^[^.#~!$@%^&*()+=,]+\.html?$", filename):
                    yield os.path.join(dirpath, filename)

    # --- FileTracer methods

    def has_dynamic_source_filename(self):
        return True

    def dynamic_source_filename(self, filename, frame):
        if frame.f_code.co_name != 'render':
            return None

        if 0:
            dump_frame(frame, label="dynamic_source_filename")
        filename = filename_for_frame(frame)
        if filename is not None:
            if filename.startswith("<"):
                # String templates have a filename of "<unknown source>", and
                # can't be reported on later, so ignore them.
                return None
            return filename
        return None

    def line_number_range(self, frame):
        assert frame.f_code.co_name == 'render'
        if 0:
            dump_frame(frame, label="line_number_range")

        render_self = frame.f_locals['self']
        if isinstance(render_self, (NodeList, Template)):
            return -1, -1

        position = position_for_node(render_self)
        if position is None:
            return -1, -1

        if SHOW_TRACING:
            print("{!r}: {}".format(render_self, position))
        s_start, s_end = position
        if isinstance(render_self, TextNode):
            first_line = render_self.s.splitlines(True)[0]
            if first_line.isspace():
                s_start += len(first_line)
        elif VerbatimNode and isinstance(render_self, VerbatimNode):
            # VerbatimNode doesn't track source the same way. s_end only points
            # to the end of the {% verbatim %} opening tag, not the entire
            # content. Adjust it to cover all of it.
            s_end += len(render_self.content)
        elif isinstance(render_self, BlockTranslateNode):
            # BlockTranslateNode has a list of text and variable tokens.
            # Get the end of the contents by looking at the last token,
            # and use its endpoint.
            last_tokens = render_self.plural or render_self.singular
            s_end = position_for_token(last_tokens[-1])[1]

        filename = filename_for_frame(frame)
        line_map = self.get_line_map(filename)
        start = get_line_number(line_map, s_start)
        end = get_line_number(line_map, s_end-1)
        if start < 0 or end < 0:
            start, end = -1, -1
        if SHOW_TRACING:
            print("line_number_range({}) -> {}".format(
                filename, (start, end)
            ))
        return start, end

    # --- FileTracer helpers

    def get_line_map(self, filename):
        """The line map for `filename`.

        A line map is a list of character offsets, indicating where each line
        in the text begins.  For example, a line map like this::

            [13, 19, 30]

        means that line 2 starts at character 13, line 3 starts at 19, etc.
        Line 1 always starts at character 0.

        """
        if filename not in self.source_map:
            template_source = read_template_source(filename)
            if 0:   # change to see the template text
                for i in range(0, len(template_source), 10):
                    print("%3d: %r" % (i, template_source[i:i+10]))
            self.source_map[filename] = make_line_map(template_source)
        return self.source_map[filename]


class FileReporter(coverage.plugin.FileReporter):
    def __init__(self, filename):
        super(FileReporter, self).__init__(filename)
        # TODO: html filenames are absolute.

        self._source = None

    def source(self):
        if self._source is None:
            self._source = read_template_source(self.filename)
        return self._source

    def lines(self):
        source_lines = set()

        if SHOW_PARSING:
            print("-------------- {}".format(self.filename))

        if django.VERSION >= (1, 9):
            lexer = Lexer(self.source())
        else:
            lexer = Lexer(self.source(), self.filename)
        tokens = lexer.tokenize()

        # Are we inside a comment?
        comment = False
        # Is this a template that extends another template?
        extends = False
        # Are we inside a block?
        inblock = False

        for token in tokens:
            if SHOW_PARSING:
                print(
                    "%10s %2d: %r" % (
                        _token_name(token.token_type),
                        token.lineno,
                        token.contents,
                    )
                )
            if token.token_type == TokenType.BLOCK:
                if token.contents == "endcomment":
                    comment = False
                    continue

            if comment:
                continue

            if token.token_type == TokenType.BLOCK:
                if token.contents.startswith("endblock"):
                    inblock = False
                elif token.contents.startswith("block"):
                    inblock = True
                    if extends:
                        continue

                if extends and not inblock:
                    # In an inheriting tempalte, ignore all tags outside of
                    # blocks.
                    continue

                if token.contents == "comment":
                    comment = True
                if token.contents.startswith("end"):
                    continue
                elif token.contents in ("else", "empty"):
                    continue
                elif token.contents.startswith("elif"):
                    # NOTE: I don't like this, I want to be able to trace elif
                    # nodes, but the Django template engine doesn't track them
                    # in a way that we can get useful information from them.
                    continue
                elif token.contents.startswith("extends"):
                    extends = True

                source_lines.add(token.lineno)

            elif token.token_type == TokenType.VAR:
                source_lines.add(token.lineno)

            elif token.token_type == TokenType.TEXT:
                if extends and not inblock:
                    continue
                # Text nodes often start with newlines, but we don't want to
                # consider that first line to be part of the text.
                lineno = token.lineno
                lines = token.contents.splitlines(True)
                num_lines = len(lines)
                if lines[0].isspace():
                    lineno += 1
                    num_lines -= 1
                source_lines.update(range(lineno, lineno+num_lines))

            if SHOW_PARSING:
                print("\t\t\tNow source_lines is: {!r}".format(source_lines))

        return source_lines


def running_sum(seq):
    total = 0
    for num in seq:
        total += num
        yield total


def make_line_map(text):
    line_lengths = [len(l) for l in text.splitlines(True)]
    line_map = list(running_sum(line_lengths))
    return line_map


def get_line_number(line_map, offset):
    """Find a line number, given a line map and a character offset."""
    for lineno, line_offset in enumerate(line_map, start=1):
        if line_offset > offset:
            return lineno
    return -1


def dump_frame(frame, label=""):
    """Dump interesting information about this frame."""
    locals = dict(frame.f_locals)
    self = locals.get('self', None)
    context = locals.get('context', None)
    if "__builtins__" in locals:
        del locals["__builtins__"]

    if label:
        label = " ( %s ) " % label
    print("-- frame --%s---------------------" % label)
    print("{}:{}:{}".format(
        os.path.basename(frame.f_code.co_filename),
        frame.f_lineno,
        type(self),
        ))
    print(locals)
    if self:
        print("self:", self.__dict__)
    if context:
        print("context:", context.__dict__)
    print("\\--")