nodes: expose line counts in node information. This would be used in full text search

rhodecode/lib/vcs/nodes.py

@@ -1,870 +1,875 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2014-2019 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 """
 Module holding everything related to vcs nodes, with vcs2 architecture.
 """
 
 import os
 import stat
 
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
 import rhodecode
 from rhodecode.config.conf import LANGUAGES_EXTENSIONS_MAP
 from rhodecode.lib.utils import safe_unicode, safe_str
 from rhodecode.lib.utils2 import md5
 from rhodecode.lib.vcs import path as vcspath
 from rhodecode.lib.vcs.backends.base import EmptyCommit, FILEMODE_DEFAULT
 from rhodecode.lib.vcs.conf.mtypes import get_mimetypes_db
 from rhodecode.lib.vcs.exceptions import NodeError, RemovedFileNodeError
 
 LARGEFILE_PREFIX = '.hglf'
 
 
 class NodeKind:
     SUBMODULE = -1
     DIR = 1
     FILE = 2
     LARGEFILE = 3
 
 
 class NodeState:
     ADDED = u'added'
     CHANGED = u'changed'
     NOT_CHANGED = u'not changed'
     REMOVED = u'removed'
 
 
 class NodeGeneratorBase(object):
     """
     Base class for removed added and changed filenodes, it's a lazy generator
     class that will create filenodes only on iteration or call
 
     The len method doesn't need to create filenodes at all
     """
 
     def __init__(self, current_paths, cs):
         self.cs = cs
         self.current_paths = current_paths
 
     def __call__(self):
         return [n for n in self]
 
     def __getslice__(self, i, j):
         for p in self.current_paths[i:j]:
             yield self.cs.get_node(p)
 
     def __len__(self):
         return len(self.current_paths)
 
     def __iter__(self):
         for p in self.current_paths:
             yield self.cs.get_node(p)
 
 
 class AddedFileNodesGenerator(NodeGeneratorBase):
     """
     Class holding added files for current commit
     """
 
 
 class ChangedFileNodesGenerator(NodeGeneratorBase):
     """
     Class holding changed files for current commit
     """
 
 
 class RemovedFileNodesGenerator(NodeGeneratorBase):
     """
     Class holding removed files for current commit
     """
     def __iter__(self):
         for p in self.current_paths:
             yield RemovedFileNode(path=p)
 
     def __getslice__(self, i, j):
         for p in self.current_paths[i:j]:
             yield RemovedFileNode(path=p)
 
 
 class Node(object):
     """
     Simplest class representing file or directory on repository.  SCM backends
     should use ``FileNode`` and ``DirNode`` subclasses rather than ``Node``
     directly.
 
     Node's ``path`` cannot start with slash as we operate on *relative* paths
     only. Moreover, every single node is identified by the ``path`` attribute,
     so it cannot end with slash, too. Otherwise, path could lead to mistakes.
     """
     RTLO_MARKER = u"\u202E"  # RTLO marker allows swapping text, and certain
                              # security attacks could be used with this
     commit = None
 
     def __init__(self, path, kind):
         self._validate_path(path)  # can throw exception if path is invalid
         self.path = safe_str(path.rstrip('/'))  # we store paths as str
         if path == '' and kind != NodeKind.DIR:
             raise NodeError("Only DirNode and its subclasses may be "
                             "initialized with empty path")
         self.kind = kind
 
         if self.is_root() and not self.is_dir():
             raise NodeError("Root node cannot be FILE kind")
 
     def _validate_path(self, path):
         if path.startswith('/'):
             raise NodeError(
                 "Cannot initialize Node objects with slash at "
                 "the beginning as only relative paths are supported. "
                 "Got %s" % (path,))
 
     @LazyProperty
     def parent(self):
         parent_path = self.get_parent_path()
         if parent_path:
             if self.commit:
                 return self.commit.get_node(parent_path)
             return DirNode(parent_path)
         return None
 
     @LazyProperty
     def unicode_path(self):
         return safe_unicode(self.path)
 
     @LazyProperty
     def has_rtlo(self):
         """Detects if a path has right-to-left-override marker"""
         return self.RTLO_MARKER in self.unicode_path
 
     @LazyProperty
     def unicode_path_safe(self):
         """
         Special SAFE representation of path without the right-to-left-override.
         This should be only used for "showing" the file, cannot be used for any
         urls etc.
         """
         return safe_unicode(self.path).replace(self.RTLO_MARKER, '')
 
     @LazyProperty
     def dir_path(self):
         """
         Returns name of the directory from full path of this vcs node. Empty
         string is returned if there's no directory in the path
         """
         _parts = self.path.rstrip('/').rsplit('/', 1)
         if len(_parts) == 2:
             return safe_unicode(_parts[0])
         return u''
 
     @LazyProperty
     def name(self):
         """
         Returns name of the node so if its path
         then only last part is returned.
         """
         return safe_unicode(self.path.rstrip('/').split('/')[-1])
 
     @property
     def kind(self):
         return self._kind
 
     @kind.setter
     def kind(self, kind):
         if hasattr(self, '_kind'):
             raise NodeError("Cannot change node's kind")
         else:
             self._kind = kind
             # Post setter check (path's trailing slash)
             if self.path.endswith('/'):
                 raise NodeError("Node's path cannot end with slash")
 
     def __cmp__(self, other):
         """
         Comparator using name of the node, needed for quick list sorting.
         """
 
         kind_cmp = cmp(self.kind, other.kind)
         if kind_cmp:
             if isinstance(self, SubModuleNode):
                 # we make submodules equal to dirnode for "sorting" purposes
                 return NodeKind.DIR
             return kind_cmp
         return cmp(self.name, other.name)
 
     def __eq__(self, other):
         for attr in ['name', 'path', 'kind']:
             if getattr(self, attr) != getattr(other, attr):
                 return False
         if self.is_file():
             if self.content != other.content:
                 return False
         else:
             # For DirNode's check without entering each dir
             self_nodes_paths = list(sorted(n.path for n in self.nodes))
             other_nodes_paths = list(sorted(n.path for n in self.nodes))
             if self_nodes_paths != other_nodes_paths:
                 return False
         return True
 
     def __ne__(self, other):
         return not self.__eq__(other)
 
     def __repr__(self):
         return '<%s %r>' % (self.__class__.__name__, self.path)
 
     def __str__(self):
         return self.__repr__()
 
     def __unicode__(self):
         return self.name
 
     def get_parent_path(self):
         """
         Returns node's parent path or empty string if node is root.
         """
         if self.is_root():
             return ''
         return vcspath.dirname(self.path.rstrip('/')) + '/'
 
     def is_file(self):
         """
         Returns ``True`` if node's kind is ``NodeKind.FILE``, ``False``
         otherwise.
         """
         return self.kind == NodeKind.FILE
 
     def is_dir(self):
         """
         Returns ``True`` if node's kind is ``NodeKind.DIR``, ``False``
         otherwise.
         """
         return self.kind == NodeKind.DIR
 
     def is_root(self):
         """
         Returns ``True`` if node is a root node and ``False`` otherwise.
         """
         return self.kind == NodeKind.DIR and self.path == ''
 
     def is_submodule(self):
         """
         Returns ``True`` if node's kind is ``NodeKind.SUBMODULE``, ``False``
         otherwise.
         """
         return self.kind == NodeKind.SUBMODULE
 
     def is_largefile(self):
         """
         Returns ``True`` if node's kind is ``NodeKind.LARGEFILE``, ``False``
         otherwise
         """
         return self.kind == NodeKind.LARGEFILE
 
     def is_link(self):
         if self.commit:
             return self.commit.is_link(self.path)
         return False
 
     @LazyProperty
     def added(self):
         return self.state is NodeState.ADDED
 
     @LazyProperty
     def changed(self):
         return self.state is NodeState.CHANGED
 
     @LazyProperty
     def not_changed(self):
         return self.state is NodeState.NOT_CHANGED
 
     @LazyProperty
     def removed(self):
         return self.state is NodeState.REMOVED
 
 
 class FileNode(Node):
     """
     Class representing file nodes.
 
     :attribute: path: path to the node, relative to repository's root
     :attribute: content: if given arbitrary sets content of the file
     :attribute: commit: if given, first time content is accessed, callback
     :attribute: mode: stat mode for a node. Default is `FILEMODE_DEFAULT`.
     """
     _filter_pre_load = []
 
     def __init__(self, path, content=None, commit=None, mode=None, pre_load=None):
         """
         Only one of ``content`` and ``commit`` may be given. Passing both
         would raise ``NodeError`` exception.
 
         :param path: relative path to the node
         :param content: content may be passed to constructor
         :param commit: if given, will use it to lazily fetch content
         :param mode: ST_MODE (i.e. 0100644)
         """
         if content and commit:
             raise NodeError("Cannot use both content and commit")
         super(FileNode, self).__init__(path, kind=NodeKind.FILE)
         self.commit = commit
         self._content = content
         self._mode = mode or FILEMODE_DEFAULT
 
         self._set_bulk_properties(pre_load)
 
     def _set_bulk_properties(self, pre_load):
         if not pre_load:
             return
         pre_load = [entry for entry in pre_load
                     if entry not in self._filter_pre_load]
         if not pre_load:
             return
 
         for attr_name in pre_load:
             result = getattr(self, attr_name)
             if callable(result):
                 result = result()
             self.__dict__[attr_name] = result
 
     @LazyProperty
     def mode(self):
         """
         Returns lazily mode of the FileNode. If `commit` is not set, would
         use value given at initialization or `FILEMODE_DEFAULT` (default).
         """
         if self.commit:
             mode = self.commit.get_file_mode(self.path)
         else:
             mode = self._mode
         return mode
 
     @LazyProperty
     def raw_bytes(self):
         """
         Returns lazily the raw bytes of the FileNode.
         """
         if self.commit:
             if self._content is None:
                 self._content = self.commit.get_file_content(self.path)
             content = self._content
         else:
             content = self._content
         return content
 
     def stream_bytes(self):
         """
         Returns an iterator that will stream the content of the file directly from
         vcsserver without loading it to memory.
         """
         if self.commit:
             return self.commit.get_file_content_streamed(self.path)
         raise NodeError("Cannot retrieve stream_bytes without related commit attribute")
 
     @LazyProperty
     def md5(self):
         """
         Returns md5 of the file node.
         """
         return md5(self.raw_bytes)
 
     def metadata_uncached(self):
         """
         Returns md5, binary flag of the file node, without any cache usage.
         """
 
         content = self.content_uncached()
 
         is_binary = content and '\0' in content
         size = 0
         if content:
             size = len(content)
 
         return is_binary, md5(content), size, content
 
     def content_uncached(self):
         """
         Returns lazily content of the FileNode. If possible, would try to
         decode content from UTF-8.
         """
         if self.commit:
             content = self.commit.get_file_content(self.path)
         else:
             content = self._content
         return content
 
     @LazyProperty
     def content(self):
         """
         Returns lazily content of the FileNode. If possible, would try to
         decode content from UTF-8.
         """
         content = self.raw_bytes
 
         if self.is_binary:
             return content
         return safe_unicode(content)
 
     @LazyProperty
     def size(self):
         if self.commit:
             return self.commit.get_file_size(self.path)
         raise NodeError(
             "Cannot retrieve size of the file without related "
             "commit attribute")
 
     @LazyProperty
     def message(self):
         if self.commit:
             return self.last_commit.message
         raise NodeError(
             "Cannot retrieve message of the file without related "
             "commit attribute")
 
     @LazyProperty
     def last_commit(self):
         if self.commit:
             pre_load = ["author", "date", "message", "parents"]
             return self.commit.get_path_commit(self.path, pre_load=pre_load)
         raise NodeError(
             "Cannot retrieve last commit of the file without "
             "related commit attribute")
 
     def get_mimetype(self):
         """
         Mimetype is calculated based on the file's content. If ``_mimetype``
         attribute is available, it will be returned (backends which store
         mimetypes or can easily recognize them, should set this private
         attribute to indicate that type should *NOT* be calculated).
         """
 
         if hasattr(self, '_mimetype'):
             if (isinstance(self._mimetype, (tuple, list,)) and
                     len(self._mimetype) == 2):
                 return self._mimetype
             else:
                 raise NodeError('given _mimetype attribute must be an 2 '
                                 'element list or tuple')
 
         db = get_mimetypes_db()
         mtype, encoding = db.guess_type(self.name)
 
         if mtype is None:
             if self.is_binary:
                 mtype = 'application/octet-stream'
                 encoding = None
             else:
                 mtype = 'text/plain'
                 encoding = None
 
                 # try with pygments
                 try:
                     from pygments.lexers import get_lexer_for_filename
                     mt = get_lexer_for_filename(self.name).mimetypes
                 except Exception:
                     mt = None
 
                 if mt:
                     mtype = mt[0]
 
         return mtype, encoding
 
     @LazyProperty
     def mimetype(self):
         """
         Wrapper around full mimetype info. It returns only type of fetched
         mimetype without the encoding part. use get_mimetype function to fetch
         full set of (type,encoding)
         """
         return self.get_mimetype()[0]
 
     @LazyProperty
     def mimetype_main(self):
         return self.mimetype.split('/')[0]
 
     @classmethod
     def get_lexer(cls, filename, content=None):
         from pygments import lexers
 
         extension = filename.split('.')[-1]
         lexer = None
 
         try:
             lexer = lexers.guess_lexer_for_filename(
                 filename, content, stripnl=False)
         except lexers.ClassNotFound:
             lexer = None
 
         # try our EXTENSION_MAP
         if not lexer:
             try:
                 lexer_class = LANGUAGES_EXTENSIONS_MAP.get(extension)
                 if lexer_class:
                     lexer = lexers.get_lexer_by_name(lexer_class[0])
             except lexers.ClassNotFound:
                 lexer = None
 
         if not lexer:
             lexer = lexers.TextLexer(stripnl=False)
 
         return lexer
 
     @LazyProperty
     def lexer(self):
         """
         Returns pygment's lexer class. Would try to guess lexer taking file's
         content, name and mimetype.
         """
         return self.get_lexer(self.name, self.content)
 
     @LazyProperty
     def lexer_alias(self):
         """
         Returns first alias of the lexer guessed for this file.
         """
         return self.lexer.aliases[0]
 
     @LazyProperty
     def history(self):
         """
         Returns a list of commit for this file in which the file was changed
         """
         if self.commit is None:
             raise NodeError('Unable to get commit for this FileNode')
         return self.commit.get_path_history(self.path)
 
     @LazyProperty
     def annotate(self):
         """
         Returns a list of three element tuples with lineno, commit and line
         """
         if self.commit is None:
             raise NodeError('Unable to get commit for this FileNode')
         pre_load = ["author", "date", "message", "parents"]
         return self.commit.get_file_annotate(self.path, pre_load=pre_load)
 
     @LazyProperty
     def state(self):
         if not self.commit:
             raise NodeError(
                 "Cannot check state of the node if it's not "
                 "linked with commit")
         elif self.path in (node.path for node in self.commit.added):
             return NodeState.ADDED
         elif self.path in (node.path for node in self.commit.changed):
             return NodeState.CHANGED
         else:
             return NodeState.NOT_CHANGED
 
     @LazyProperty
     def is_binary(self):
         """
         Returns True if file has binary content.
         """
         if self.commit:
             return self.commit.is_node_binary(self.path)
         else:
             raw_bytes = self._content
             return raw_bytes and '\0' in raw_bytes
 
     @LazyProperty
     def extension(self):
         """Returns filenode extension"""
         return self.name.split('.')[-1]
 
     @property
     def is_executable(self):
         """
         Returns ``True`` if file has executable flag turned on.
         """
         return bool(self.mode & stat.S_IXUSR)
 
     def get_largefile_node(self):
         """
         Try to return a Mercurial FileNode from this node. It does internal
         checks inside largefile store, if that file exist there it will
         create special instance of LargeFileNode which can get content from
         LF store.
         """
         if self.commit:
             return self.commit.get_largefile_node(self.path)
 
+    def count_lines(self, content, count_empty=False):
+
+        if count_empty:
+            all_lines = 0
+            empty_lines = 0
+            for line in content.splitlines(True):
+                if line == '\n':
+                    empty_lines += 1
+                all_lines += 1
+
+            return all_lines, all_lines - empty_lines
+        else:
+            # fast method
+            empty_lines = all_lines = content.count('\n')
+            if all_lines == 0 and content:
+                # one-line without a newline
+                empty_lines = all_lines = 1
+
+        return all_lines, empty_lines
+
     def lines(self, count_empty=False):
         all_lines, empty_lines = 0, 0
 
         if not self.is_binary:
             content = self.content
-            if count_empty:
-                all_lines = 0
-                empty_lines = 0
-                for line in content.splitlines(True):
-                    if line == '\n':
-                        empty_lines += 1
-                    all_lines += 1
-
-                return all_lines, all_lines - empty_lines
-            else:
-                # fast method
-                empty_lines = all_lines = content.count('\n')
-                if all_lines == 0 and content:
-                    # one-line without a newline
-                    empty_lines = all_lines = 1
-
+            all_lines, empty_lines = self.count_lines(content, count_empty=count_empty)
         return all_lines, empty_lines
 
     def __repr__(self):
         return '<%s %r @ %s>' % (self.__class__.__name__, self.path,
                                  getattr(self.commit, 'short_id', ''))
 
 
 class RemovedFileNode(FileNode):
     """
     Dummy FileNode class - trying to access any public attribute except path,
     name, kind or state (or methods/attributes checking those two) would raise
     RemovedFileNodeError.
     """
     ALLOWED_ATTRIBUTES = [
         'name', 'path', 'state', 'is_root', 'is_file', 'is_dir', 'kind',
         'added', 'changed', 'not_changed', 'removed'
     ]
 
     def __init__(self, path):
         """
         :param path: relative path to the node
         """
         super(RemovedFileNode, self).__init__(path=path)
 
     def __getattribute__(self, attr):
         if attr.startswith('_') or attr in RemovedFileNode.ALLOWED_ATTRIBUTES:
             return super(RemovedFileNode, self).__getattribute__(attr)
         raise RemovedFileNodeError(
             "Cannot access attribute %s on RemovedFileNode" % attr)
 
     @LazyProperty
     def state(self):
         return NodeState.REMOVED
 
 
 class DirNode(Node):
     """
     DirNode stores list of files and directories within this node.
     Nodes may be used standalone but within repository context they
     lazily fetch data within same repository's commit.
     """
 
     def __init__(self, path, nodes=(), commit=None):
         """
         Only one of ``nodes`` and ``commit`` may be given. Passing both
         would raise ``NodeError`` exception.
 
         :param path: relative path to the node
         :param nodes: content may be passed to constructor
         :param commit: if given, will use it to lazily fetch content
         """
         if nodes and commit:
             raise NodeError("Cannot use both nodes and commit")
         super(DirNode, self).__init__(path, NodeKind.DIR)
         self.commit = commit
         self._nodes = nodes
 
     @LazyProperty
     def content(self):
         raise NodeError(
             "%s represents a dir and has no `content` attribute" % self)
 
     @LazyProperty
     def nodes(self):
         if self.commit:
             nodes = self.commit.get_nodes(self.path)
         else:
             nodes = self._nodes
         self._nodes_dict = dict((node.path, node) for node in nodes)
         return sorted(nodes)
 
     @LazyProperty
     def files(self):
         return sorted((node for node in self.nodes if node.is_file()))
 
     @LazyProperty
     def dirs(self):
         return sorted((node for node in self.nodes if node.is_dir()))
 
     def __iter__(self):
         for node in self.nodes:
             yield node
 
     def get_node(self, path):
         """
         Returns node from within this particular ``DirNode``, so it is now
         allowed to fetch, i.e. node located at 'docs/api/index.rst' from node
         'docs'. In order to access deeper nodes one must fetch nodes between
         them first - this would work::
 
            docs = root.get_node('docs')
            docs.get_node('api').get_node('index.rst')
 
         :param: path - relative to the current node
 
         .. note::
            To access lazily (as in example above) node have to be initialized
            with related commit object - without it node is out of
            context and may know nothing about anything else than nearest
            (located at same level) nodes.
         """
         try:
             path = path.rstrip('/')
             if path == '':
                 raise NodeError("Cannot retrieve node without path")
             self.nodes  # access nodes first in order to set _nodes_dict
             paths = path.split('/')
             if len(paths) == 1:
                 if not self.is_root():
                     path = '/'.join((self.path, paths[0]))
                 else:
                     path = paths[0]
                 return self._nodes_dict[path]
             elif len(paths) > 1:
                 if self.commit is None:
                     raise NodeError("Cannot access deeper nodes without commit")
                 else:
                     path1, path2 = paths[0], '/'.join(paths[1:])
                     return self.get_node(path1).get_node(path2)
             else:
                 raise KeyError
         except KeyError:
             raise NodeError("Node does not exist at %s" % path)
 
     @LazyProperty
     def state(self):
         raise NodeError("Cannot access state of DirNode")
 
     @LazyProperty
     def size(self):
         size = 0
         for root, dirs, files in self.commit.walk(self.path):
             for f in files:
                 size += f.size
 
         return size
 
     @LazyProperty
     def last_commit(self):
         if self.commit:
             pre_load = ["author", "date", "message", "parents"]
             return self.commit.get_path_commit(self.path, pre_load=pre_load)
         raise NodeError(
             "Cannot retrieve last commit of the file without "
             "related commit attribute")
 
     def __repr__(self):
         return '<%s %r @ %s>' % (self.__class__.__name__, self.path,
                                  getattr(self.commit, 'short_id', ''))
 
 
 class RootNode(DirNode):
     """
     DirNode being the root node of the repository.
     """
 
     def __init__(self, nodes=(), commit=None):
         super(RootNode, self).__init__(path='', nodes=nodes, commit=commit)
 
     def __repr__(self):
         return '<%s>' % self.__class__.__name__
 
 
 class SubModuleNode(Node):
     """
     represents a SubModule of Git or SubRepo of Mercurial
     """
     is_binary = False
     size = 0
 
     def __init__(self, name, url=None, commit=None, alias=None):
         self.path = name
         self.kind = NodeKind.SUBMODULE
         self.alias = alias
 
         # we have to use EmptyCommit here since this can point to svn/git/hg
         # submodules we cannot get from repository
         self.commit = EmptyCommit(str(commit), alias=alias)
         self.url = url or self._extract_submodule_url()
 
     def __repr__(self):
         return '<%s %r @ %s>' % (self.__class__.__name__, self.path,
                                  getattr(self.commit, 'short_id', ''))
 
     def _extract_submodule_url(self):
         # TODO: find a way to parse gits submodule file and extract the
         # linking URL
         return self.path
 
     @LazyProperty
     def name(self):
         """
         Returns name of the node so if its path
         then only last part is returned.
         """
         org = safe_unicode(self.path.rstrip('/').split('/')[-1])
         return u'%s @ %s' % (org, self.commit.short_id)
 
 
 class LargeFileNode(FileNode):
 
     def __init__(self, path, url=None, commit=None, alias=None, org_path=None):
         self.path = path
         self.org_path = org_path
         self.kind = NodeKind.LARGEFILE
         self.alias = alias
 
     def _validate_path(self, path):
         """
         we override check since the LargeFileNode path is system absolute
         """
         pass
 
     def __repr__(self):
         return '<%s %r>' % (self.__class__.__name__, self.path)
 
     @LazyProperty
     def size(self):
         return os.stat(self.path).st_size
 
     @LazyProperty
     def raw_bytes(self):
         with open(self.path, 'rb') as f:
             content = f.read()
         return content
 
     @LazyProperty
     def name(self):
         """
         Overwrites name to be the org lf path
         """
         return self.org_path
 
     def stream_bytes(self):
         with open(self.path, 'rb') as stream:
             while True:
                 data = stream.read(16 * 1024)
                 if not data:
                     break
                 yield data

rhodecode/model/scm.py

@@ -1,1008 +1,1014 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2010-2019 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 """
 Scm model for RhodeCode
 """
 
 import os.path
 import traceback
 import logging
 import cStringIO
 
 from sqlalchemy import func
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
 import rhodecode
 from rhodecode.lib.vcs import get_backend
 from rhodecode.lib.vcs.exceptions import RepositoryError, NodeNotChangedError
 from rhodecode.lib.vcs.nodes import FileNode
 from rhodecode.lib.vcs.backends.base import EmptyCommit
 from rhodecode.lib import helpers as h, rc_cache
 from rhodecode.lib.auth import (
     HasRepoPermissionAny, HasRepoGroupPermissionAny,
     HasUserGroupPermissionAny)
 from rhodecode.lib.exceptions import NonRelativePathError, IMCCommitError
 from rhodecode.lib import hooks_utils
 from rhodecode.lib.utils import (
     get_filesystem_repos, make_db_config)
 from rhodecode.lib.utils2 import (safe_str, safe_unicode)
 from rhodecode.lib.system_info import get_system_info
 from rhodecode.model import BaseModel
 from rhodecode.model.db import (
     Repository, CacheKey, UserFollowing, UserLog, User, RepoGroup,
     PullRequest)
 from rhodecode.model.settings import VcsSettingsModel
 from rhodecode.model.validation_schema.validators import url_validator, InvalidCloneUrl
 
 log = logging.getLogger(__name__)
 
 
 class UserTemp(object):
     def __init__(self, user_id):
         self.user_id = user_id
 
     def __repr__(self):
         return "<%s('id:%s')>" % (self.__class__.__name__, self.user_id)
 
 
 class RepoTemp(object):
     def __init__(self, repo_id):
         self.repo_id = repo_id
 
     def __repr__(self):
         return "<%s('id:%s')>" % (self.__class__.__name__, self.repo_id)
 
 
 class SimpleCachedRepoList(object):
     """
     Lighter version of of iteration of repos without the scm initialisation,
     and with cache usage
     """
     def __init__(self, db_repo_list, repos_path, order_by=None, perm_set=None):
         self.db_repo_list = db_repo_list
         self.repos_path = repos_path
         self.order_by = order_by
         self.reversed = (order_by or '').startswith('-')
         if not perm_set:
             perm_set = ['repository.read', 'repository.write',
                         'repository.admin']
         self.perm_set = perm_set
 
     def __len__(self):
         return len(self.db_repo_list)
 
     def __repr__(self):
         return '<%s (%s)>' % (self.__class__.__name__, self.__len__())
 
     def __iter__(self):
         for dbr in self.db_repo_list:
             # check permission at this level
             has_perm = HasRepoPermissionAny(*self.perm_set)(
                 dbr.repo_name, 'SimpleCachedRepoList check')
             if not has_perm:
                 continue
 
             tmp_d = {
                 'name': dbr.repo_name,
                 'dbrepo': dbr.get_dict(),
                 'dbrepo_fork': dbr.fork.get_dict() if dbr.fork else {}
             }
             yield tmp_d
 
 
 class _PermCheckIterator(object):
 
     def __init__(
             self, obj_list, obj_attr, perm_set, perm_checker,
             extra_kwargs=None):
         """
         Creates iterator from given list of objects, additionally
         checking permission for them from perm_set var
 
         :param obj_list: list of db objects
         :param obj_attr: attribute of object to pass into perm_checker
         :param perm_set: list of permissions to check
         :param perm_checker: callable to check permissions against
         """
         self.obj_list = obj_list
         self.obj_attr = obj_attr
         self.perm_set = perm_set
         self.perm_checker = perm_checker
         self.extra_kwargs = extra_kwargs or {}
 
     def __len__(self):
         return len(self.obj_list)
 
     def __repr__(self):
         return '<%s (%s)>' % (self.__class__.__name__, self.__len__())
 
     def __iter__(self):
         checker = self.perm_checker(*self.perm_set)
         for db_obj in self.obj_list:
             # check permission at this level
             name = getattr(db_obj, self.obj_attr, None)
             if not checker(name, self.__class__.__name__, **self.extra_kwargs):
                 continue
 
             yield db_obj
 
 
 class RepoList(_PermCheckIterator):
 
     def __init__(self, db_repo_list, perm_set=None, extra_kwargs=None):
         if not perm_set:
             perm_set = [
                 'repository.read', 'repository.write', 'repository.admin']
 
         super(RepoList, self).__init__(
             obj_list=db_repo_list,
             obj_attr='repo_name', perm_set=perm_set,
             perm_checker=HasRepoPermissionAny,
             extra_kwargs=extra_kwargs)
 
 
 class RepoGroupList(_PermCheckIterator):
 
     def __init__(self, db_repo_group_list, perm_set=None, extra_kwargs=None):
         if not perm_set:
             perm_set = ['group.read', 'group.write', 'group.admin']
 
         super(RepoGroupList, self).__init__(
             obj_list=db_repo_group_list,
             obj_attr='group_name', perm_set=perm_set,
             perm_checker=HasRepoGroupPermissionAny,
             extra_kwargs=extra_kwargs)
 
 
 class UserGroupList(_PermCheckIterator):
 
     def __init__(self, db_user_group_list, perm_set=None, extra_kwargs=None):
         if not perm_set:
             perm_set = ['usergroup.read', 'usergroup.write', 'usergroup.admin']
 
         super(UserGroupList, self).__init__(
             obj_list=db_user_group_list,
             obj_attr='users_group_name', perm_set=perm_set,
             perm_checker=HasUserGroupPermissionAny,
             extra_kwargs=extra_kwargs)
 
 
 class ScmModel(BaseModel):
     """
     Generic Scm Model
     """
 
     @LazyProperty
     def repos_path(self):
         """
         Gets the repositories root path from database
         """
 
         settings_model = VcsSettingsModel(sa=self.sa)
         return settings_model.get_repos_location()
 
     def repo_scan(self, repos_path=None):
         """
         Listing of repositories in given path. This path should not be a
         repository itself. Return a dictionary of repository objects
 
         :param repos_path: path to directory containing repositories
         """
 
         if repos_path is None:
             repos_path = self.repos_path
 
         log.info('scanning for repositories in %s', repos_path)
 
         config = make_db_config()
         config.set('extensions', 'largefiles', '')
         repos = {}
 
         for name, path in get_filesystem_repos(repos_path, recursive=True):
             # name need to be decomposed and put back together using the /
             # since this is internal storage separator for rhodecode
             name = Repository.normalize_repo_name(name)
 
             try:
                 if name in repos:
                     raise RepositoryError('Duplicate repository name %s '
                                           'found in %s' % (name, path))
                 elif path[0] in rhodecode.BACKENDS:
                     backend = get_backend(path[0])
                     repos[name] = backend(path[1], config=config,
                                           with_wire={"cache": False})
             except OSError:
                 continue
         log.debug('found %s paths with repositories', len(repos))
         return repos
 
     def get_repos(self, all_repos=None, sort_key=None):
         """
         Get all repositories from db and for each repo create it's
         backend instance and fill that backed with information from database
 
         :param all_repos: list of repository names as strings
             give specific repositories list, good for filtering
 
         :param sort_key: initial sorting of repositories
         """
         if all_repos is None:
             all_repos = self.sa.query(Repository)\
                 .filter(Repository.group_id == None)\
                 .order_by(func.lower(Repository.repo_name)).all()
         repo_iter = SimpleCachedRepoList(
             all_repos, repos_path=self.repos_path, order_by=sort_key)
         return repo_iter
 
     def get_repo_groups(self, all_groups=None):
         if all_groups is None:
             all_groups = RepoGroup.query()\
                 .filter(RepoGroup.group_parent_id == None).all()
         return [x for x in RepoGroupList(all_groups)]
 
     def mark_for_invalidation(self, repo_name, delete=False):
         """
         Mark caches of this repo invalid in the database. `delete` flag
         removes the cache entries
 
         :param repo_name: the repo_name for which caches should be marked
             invalid, or deleted
         :param delete: delete the entry keys instead of setting bool
             flag on them, and also purge caches used by the dogpile
         """
         repo = Repository.get_by_repo_name(repo_name)
 
         if repo:
             invalidation_namespace = CacheKey.REPO_INVALIDATION_NAMESPACE.format(
                 repo_id=repo.repo_id)
             CacheKey.set_invalidate(invalidation_namespace, delete=delete)
 
             repo_id = repo.repo_id
             config = repo._config
             config.set('extensions', 'largefiles', '')
             repo.update_commit_cache(config=config, cs_cache=None)
             if delete:
                 cache_namespace_uid = 'cache_repo.{}'.format(repo_id)
                 rc_cache.clear_cache_namespace('cache_repo', cache_namespace_uid)
 
     def toggle_following_repo(self, follow_repo_id, user_id):
 
         f = self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_repo_id == follow_repo_id)\
             .filter(UserFollowing.user_id == user_id).scalar()
 
         if f is not None:
             try:
                 self.sa.delete(f)
                 return
             except Exception:
                 log.error(traceback.format_exc())
                 raise
 
         try:
             f = UserFollowing()
             f.user_id = user_id
             f.follows_repo_id = follow_repo_id
             self.sa.add(f)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def toggle_following_user(self, follow_user_id, user_id):
         f = self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_user_id == follow_user_id)\
             .filter(UserFollowing.user_id == user_id).scalar()
 
         if f is not None:
             try:
                 self.sa.delete(f)
                 return
             except Exception:
                 log.error(traceback.format_exc())
                 raise
 
         try:
             f = UserFollowing()
             f.user_id = user_id
             f.follows_user_id = follow_user_id
             self.sa.add(f)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def is_following_repo(self, repo_name, user_id, cache=False):
         r = self.sa.query(Repository)\
             .filter(Repository.repo_name == repo_name).scalar()
 
         f = self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_repository == r)\
             .filter(UserFollowing.user_id == user_id).scalar()
 
         return f is not None
 
     def is_following_user(self, username, user_id, cache=False):
         u = User.get_by_username(username)
 
         f = self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_user == u)\
             .filter(UserFollowing.user_id == user_id).scalar()
 
         return f is not None
 
     def get_followers(self, repo):
         repo = self._get_repo(repo)
 
         return self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_repository == repo).count()
 
     def get_forks(self, repo):
         repo = self._get_repo(repo)
         return self.sa.query(Repository)\
             .filter(Repository.fork == repo).count()
 
     def get_pull_requests(self, repo):
         repo = self._get_repo(repo)
         return self.sa.query(PullRequest)\
             .filter(PullRequest.target_repo == repo)\
             .filter(PullRequest.status != PullRequest.STATUS_CLOSED).count()
 
     def mark_as_fork(self, repo, fork, user):
         repo = self._get_repo(repo)
         fork = self._get_repo(fork)
         if fork and repo.repo_id == fork.repo_id:
             raise Exception("Cannot set repository as fork of itself")
 
         if fork and repo.repo_type != fork.repo_type:
             raise RepositoryError(
                 "Cannot set repository as fork of repository with other type")
 
         repo.fork = fork
         self.sa.add(repo)
         return repo
 
     def pull_changes(self, repo, username, remote_uri=None, validate_uri=True):
         dbrepo = self._get_repo(repo)
         remote_uri = remote_uri or dbrepo.clone_uri
         if not remote_uri:
             raise Exception("This repository doesn't have a clone uri")
 
         repo = dbrepo.scm_instance(cache=False)
         repo.config.clear_section('hooks')
 
         try:
             # NOTE(marcink): add extra validation so we skip invalid urls
             # this is due this tasks can be executed via scheduler without
             # proper validation of remote_uri
             if validate_uri:
                 config = make_db_config(clear_session=False)
                 url_validator(remote_uri, dbrepo.repo_type, config)
         except InvalidCloneUrl:
             raise
 
         repo_name = dbrepo.repo_name
         try:
             # TODO: we need to make sure those operations call proper hooks !
             repo.fetch(remote_uri)
 
             self.mark_for_invalidation(repo_name)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def push_changes(self, repo, username, remote_uri=None, validate_uri=True):
         dbrepo = self._get_repo(repo)
         remote_uri = remote_uri or dbrepo.push_uri
         if not remote_uri:
             raise Exception("This repository doesn't have a clone uri")
 
         repo = dbrepo.scm_instance(cache=False)
         repo.config.clear_section('hooks')
 
         try:
             # NOTE(marcink): add extra validation so we skip invalid urls
             # this is due this tasks can be executed via scheduler without
             # proper validation of remote_uri
             if validate_uri:
                 config = make_db_config(clear_session=False)
                 url_validator(remote_uri, dbrepo.repo_type, config)
         except InvalidCloneUrl:
             raise
 
         try:
             repo.push(remote_uri)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def commit_change(self, repo, repo_name, commit, user, author, message,
                       content, f_path):
         """
         Commits changes
 
         :param repo: SCM instance
 
         """
         user = self._get_user(user)
 
         # decoding here will force that we have proper encoded values
         # in any other case this will throw exceptions and deny commit
         content = safe_str(content)
         path = safe_str(f_path)
         # message and author needs to be unicode
         # proper backend should then translate that into required type
         message = safe_unicode(message)
         author = safe_unicode(author)
         imc = repo.in_memory_commit
         imc.change(FileNode(path, content, mode=commit.get_file_mode(f_path)))
         try:
             # TODO: handle pre-push action !
             tip = imc.commit(
                 message=message, author=author, parents=[commit],
                 branch=commit.branch)
         except Exception as e:
             log.error(traceback.format_exc())
             raise IMCCommitError(str(e))
         finally:
             # always clear caches, if commit fails we want fresh object also
             self.mark_for_invalidation(repo_name)
 
         # We trigger the post-push action
         hooks_utils.trigger_post_push_hook(
             username=user.username, action='push_local', hook_type='post_push',
             repo_name=repo_name, repo_alias=repo.alias, commit_ids=[tip.raw_id])
         return tip
 
     def _sanitize_path(self, f_path):
         if f_path.startswith('/') or f_path.startswith('./') or '../' in f_path:
             raise NonRelativePathError('%s is not an relative path' % f_path)
         if f_path:
             f_path = os.path.normpath(f_path)
         return f_path
 
     def get_dirnode_metadata(self, request, commit, dir_node):
         if not dir_node.is_dir():
             return []
 
         data = []
         for node in dir_node:
             if not node.is_file():
                 # we skip file-nodes
                 continue
 
             last_commit = node.last_commit
             last_commit_date = last_commit.date
             data.append({
                 'name': node.name,
                 'size': h.format_byte_size_binary(node.size),
                 'modified_at': h.format_date(last_commit_date),
                 'modified_ts': last_commit_date.isoformat(),
                 'revision': last_commit.revision,
                 'short_id': last_commit.short_id,
                 'message': h.escape(last_commit.message),
                 'author': h.escape(last_commit.author),
                 'user_profile': h.gravatar_with_user(
                     request, last_commit.author),
             })
 
         return data
 
     def get_nodes(self, repo_name, commit_id, root_path='/', flat=True,
                   extended_info=False, content=False, max_file_bytes=None):
         """
         recursive walk in root dir and return a set of all path in that dir
         based on repository walk function
 
         :param repo_name: name of repository
         :param commit_id: commit id for which to list nodes
         :param root_path: root path to list
         :param flat: return as a list, if False returns a dict with description
         :param extended_info: show additional info such as md5, binary, size etc
         :param content: add nodes content to the return data
         :param max_file_bytes: will not return file contents over this limit
 
         """
         _files = list()
         _dirs = list()
         try:
             _repo = self._get_repo(repo_name)
             commit = _repo.scm_instance().get_commit(commit_id=commit_id)
             root_path = root_path.lstrip('/')
             for __, dirs, files in commit.walk(root_path):
 
                 for f in files:
                     _content = None
                     _data = f_name = f.unicode_path
 
                     if not flat:
                         _data = {
                             "name": h.escape(f_name),
                             "type": "file",
                             }
                         if extended_info:
                             _data.update({
                                 "md5": f.md5,
                                 "binary": f.is_binary,
                                 "size": f.size,
                                 "extension": f.extension,
                                 "mimetype": f.mimetype,
                                 "lines": f.lines()[0]
                             })
 
                         if content:
                             over_size_limit = (max_file_bytes is not None
                                                and f.size > max_file_bytes)
                             full_content = None
                             if not f.is_binary and not over_size_limit:
                                 full_content = safe_str(f.content)
 
                             _data.update({
                                 "content": full_content,
                             })
                     _files.append(_data)
 
                 for d in dirs:
                     _data = d_name = d.unicode_path
                     if not flat:
                         _data = {
                             "name": h.escape(d_name),
                             "type": "dir",
                             }
                     if extended_info:
                         _data.update({
                             "md5": None,
                             "binary": None,
                             "size": None,
                             "extension": None,
                         })
                     if content:
                         _data.update({
                             "content": None
                         })
                     _dirs.append(_data)
         except RepositoryError:
             log.exception("Exception in get_nodes")
             raise
 
         return _dirs, _files
 
     def get_quick_filter_nodes(self, repo_name, commit_id, root_path='/'):
         """
         Generate files for quick filter in files view
         """
 
         _files = list()
         _dirs = list()
         try:
             _repo = self._get_repo(repo_name)
             commit = _repo.scm_instance().get_commit(commit_id=commit_id)
             root_path = root_path.lstrip('/')
             for __, dirs, files in commit.walk(root_path):
 
                 for f in files:
 
                     _data = {
                         "name": h.escape(f.unicode_path),
                         "type": "file",
                         }
 
                     _files.append(_data)
 
                 for d in dirs:
 
                     _data = {
                         "name": h.escape(d.unicode_path),
                         "type": "dir",
                         }
 
                     _dirs.append(_data)
         except RepositoryError:
             log.exception("Exception in get_quick_filter_nodes")
             raise
 
         return _dirs, _files
 
     def get_node(self, repo_name, commit_id, file_path,
                  extended_info=False, content=False, max_file_bytes=None, cache=True):
         """
         retrieve single node from commit
         """
         try:
 
             _repo = self._get_repo(repo_name)
             commit = _repo.scm_instance().get_commit(commit_id=commit_id)
 
             file_node = commit.get_node(file_path)
             if file_node.is_dir():
                 raise RepositoryError('The given path is a directory')
 
             _content = None
             f_name = file_node.unicode_path
 
             file_data = {
                 "name": h.escape(f_name),
                 "type": "file",
             }
 
             if extended_info:
                 file_data.update({
                     "extension": file_node.extension,
                     "mimetype": file_node.mimetype,
                 })
 
                 if cache:
                     md5 = file_node.md5
                     is_binary = file_node.is_binary
                     size = file_node.size
                 else:
                     is_binary, md5, size, _content = file_node.metadata_uncached()
 
                 file_data.update({
                     "md5": md5,
                     "binary": is_binary,
                     "size": size,
                 })
 
             if content and cache:
                 # get content + cache
                 size = file_node.size
                 over_size_limit = (max_file_bytes is not None and size > max_file_bytes)
                 full_content = None
+                all_lines = 0
                 if not file_node.is_binary and not over_size_limit:
                     full_content = safe_unicode(file_node.content)
+                    all_lines, empty_lines = file_node.count_lines(full_content)
 
                 file_data.update({
                     "content": full_content,
+                    "lines": all_lines
                 })
             elif content:
                 # get content *without* cache
                 if _content is None:
                     is_binary, md5, size, _content = file_node.metadata_uncached()
 
                 over_size_limit = (max_file_bytes is not None and size > max_file_bytes)
                 full_content = None
+                all_lines = 0
                 if not is_binary and not over_size_limit:
                     full_content = safe_unicode(_content)
+                    all_lines, empty_lines = file_node.count_lines(full_content)
 
                 file_data.update({
                     "content": full_content,
+                    "lines": all_lines
                 })
 
         except RepositoryError:
             log.exception("Exception in get_node")
             raise
 
         return file_data
 
     def get_fts_data(self, repo_name, commit_id, root_path='/'):
         """
         Fetch node tree for usage in full text search
         """
 
         tree_info = list()
 
         try:
             _repo = self._get_repo(repo_name)
             commit = _repo.scm_instance().get_commit(commit_id=commit_id)
             root_path = root_path.lstrip('/')
             for __, dirs, files in commit.walk(root_path):
 
                 for f in files:
                     is_binary, md5, size, _content = f.metadata_uncached()
                     _data = {
                         "name": f.unicode_path,
                         "md5": md5,
                         "extension": f.extension,
                         "binary": is_binary,
                         "size": size
                     }
 
                     tree_info.append(_data)
 
         except RepositoryError:
             log.exception("Exception in get_nodes")
             raise
 
         return tree_info
 
     def create_nodes(self, user, repo, message, nodes, parent_commit=None,
                      author=None, trigger_push_hook=True):
         """
         Commits given multiple nodes into repo
 
         :param user: RhodeCode User object or user_id, the commiter
         :param repo: RhodeCode Repository object
         :param message: commit message
         :param nodes: mapping {filename:{'content':content},...}
         :param parent_commit: parent commit, can be empty than it's
            initial commit
         :param author: author of commit, cna be different that commiter
            only for git
         :param trigger_push_hook: trigger push hooks
 
         :returns: new commited commit
         """
 
         user = self._get_user(user)
         scm_instance = repo.scm_instance(cache=False)
 
         processed_nodes = []
         for f_path in nodes:
             f_path = self._sanitize_path(f_path)
             content = nodes[f_path]['content']
             f_path = safe_str(f_path)
             # decoding here will force that we have proper encoded values
             # in any other case this will throw exceptions and deny commit
             if isinstance(content, (basestring,)):
                 content = safe_str(content)
             elif isinstance(content, (file, cStringIO.OutputType,)):
                 content = content.read()
             else:
                 raise Exception('Content is of unrecognized type %s' % (
                     type(content)
                 ))
             processed_nodes.append((f_path, content))
 
         message = safe_unicode(message)
         commiter = user.full_contact
         author = safe_unicode(author) if author else commiter
 
         imc = scm_instance.in_memory_commit
 
         if not parent_commit:
             parent_commit = EmptyCommit(alias=scm_instance.alias)
 
         if isinstance(parent_commit, EmptyCommit):
             # EmptyCommit means we we're editing empty repository
             parents = None
         else:
             parents = [parent_commit]
         # add multiple nodes
         for path, content in processed_nodes:
             imc.add(FileNode(path, content=content))
         # TODO: handle pre push scenario
         tip = imc.commit(message=message,
                          author=author,
                          parents=parents,
                          branch=parent_commit.branch)
 
         self.mark_for_invalidation(repo.repo_name)
         if trigger_push_hook:
             hooks_utils.trigger_post_push_hook(
                 username=user.username, action='push_local',
                 repo_name=repo.repo_name, repo_alias=scm_instance.alias,
                 hook_type='post_push',
                 commit_ids=[tip.raw_id])
         return tip
 
     def update_nodes(self, user, repo, message, nodes, parent_commit=None,
                      author=None, trigger_push_hook=True):
         user = self._get_user(user)
         scm_instance = repo.scm_instance(cache=False)
 
         message = safe_unicode(message)
         commiter = user.full_contact
         author = safe_unicode(author) if author else commiter
 
         imc = scm_instance.in_memory_commit
 
         if not parent_commit:
             parent_commit = EmptyCommit(alias=scm_instance.alias)
 
         if isinstance(parent_commit, EmptyCommit):
             # EmptyCommit means we we're editing empty repository
             parents = None
         else:
             parents = [parent_commit]
 
         # add multiple nodes
         for _filename, data in nodes.items():
             # new filename, can be renamed from the old one, also sanitaze
             # the path for any hack around relative paths like ../../ etc.
             filename = self._sanitize_path(data['filename'])
             old_filename = self._sanitize_path(_filename)
             content = data['content']
             file_mode = data.get('mode')
             filenode = FileNode(old_filename, content=content, mode=file_mode)
             op = data['op']
             if op == 'add':
                 imc.add(filenode)
             elif op == 'del':
                 imc.remove(filenode)
             elif op == 'mod':
                 if filename != old_filename:
                     # TODO: handle renames more efficient, needs vcs lib changes
                     imc.remove(filenode)
                     imc.add(FileNode(filename, content=content, mode=file_mode))
                 else:
                     imc.change(filenode)
 
         try:
             # TODO: handle pre push scenario commit changes
             tip = imc.commit(message=message,
                              author=author,
                              parents=parents,
                              branch=parent_commit.branch)
         except NodeNotChangedError:
             raise
         except Exception as e:
             log.exception("Unexpected exception during call to imc.commit")
             raise IMCCommitError(str(e))
         finally:
             # always clear caches, if commit fails we want fresh object also
             self.mark_for_invalidation(repo.repo_name)
 
         if trigger_push_hook:
             hooks_utils.trigger_post_push_hook(
                 username=user.username, action='push_local', hook_type='post_push',
                 repo_name=repo.repo_name, repo_alias=scm_instance.alias,
                 commit_ids=[tip.raw_id])
 
         return tip
 
     def delete_nodes(self, user, repo, message, nodes, parent_commit=None,
                      author=None, trigger_push_hook=True):
         """
         Deletes given multiple nodes into `repo`
 
         :param user: RhodeCode User object or user_id, the committer
         :param repo: RhodeCode Repository object
         :param message: commit message
         :param nodes: mapping {filename:{'content':content},...}
         :param parent_commit: parent commit, can be empty than it's initial
            commit
         :param author: author of commit, cna be different that commiter only
            for git
         :param trigger_push_hook: trigger push hooks
 
         :returns: new commit after deletion
         """
 
         user = self._get_user(user)
         scm_instance = repo.scm_instance(cache=False)
 
         processed_nodes = []
         for f_path in nodes:
             f_path = self._sanitize_path(f_path)
             # content can be empty but for compatabilty it allows same dicts
             # structure as add_nodes
             content = nodes[f_path].get('content')
             processed_nodes.append((f_path, content))
 
         message = safe_unicode(message)
         commiter = user.full_contact
         author = safe_unicode(author) if author else commiter
 
         imc = scm_instance.in_memory_commit
 
         if not parent_commit:
             parent_commit = EmptyCommit(alias=scm_instance.alias)
 
         if isinstance(parent_commit, EmptyCommit):
             # EmptyCommit means we we're editing empty repository
             parents = None
         else:
             parents = [parent_commit]
         # add multiple nodes
         for path, content in processed_nodes:
             imc.remove(FileNode(path, content=content))
 
         # TODO: handle pre push scenario
         tip = imc.commit(message=message,
                          author=author,
                          parents=parents,
                          branch=parent_commit.branch)
 
         self.mark_for_invalidation(repo.repo_name)
         if trigger_push_hook:
             hooks_utils.trigger_post_push_hook(
                 username=user.username, action='push_local', hook_type='post_push',
                 repo_name=repo.repo_name, repo_alias=scm_instance.alias,
                 commit_ids=[tip.raw_id])
         return tip
 
     def strip(self, repo, commit_id, branch):
         scm_instance = repo.scm_instance(cache=False)
         scm_instance.config.clear_section('hooks')
         scm_instance.strip(commit_id, branch)
         self.mark_for_invalidation(repo.repo_name)
 
     def get_unread_journal(self):
         return self.sa.query(UserLog).count()
 
     @classmethod
     def backend_landing_ref(cls, repo_type):
         """
         Return a default landing ref based on a repository type.
         """
 
         landing_ref = {
             'hg': ('branch:default', 'default'),
             'git': ('branch:master', 'master'),
             'svn': ('rev:tip', 'latest tip'),
             'default': ('rev:tip', 'latest tip'),
         }
 
         return landing_ref.get(repo_type) or landing_ref['default']
 
     def get_repo_landing_revs(self, translator, repo=None):
         """
         Generates select option with tags branches and bookmarks (for hg only)
         grouped by type
 
         :param repo:
         """
         _ = translator
         repo = self._get_repo(repo)
 
         if repo:
             repo_type = repo.repo_type
         else:
             repo_type = 'default'
 
         default_landing_ref, landing_ref_lbl = self.backend_landing_ref(repo_type)
 
         default_ref_options = [
             [default_landing_ref, landing_ref_lbl]
         ]
         default_choices = [
             default_landing_ref
         ]
 
         if not repo:
             return default_choices, default_ref_options
 
         repo = repo.scm_instance()
 
         ref_options = [('rev:tip', 'latest tip')]
         choices = ['rev:tip']
 
         # branches
         branch_group = [(u'branch:%s' % safe_unicode(b), safe_unicode(b)) for b in repo.branches]
         if not branch_group:
             # new repo, or without maybe a branch?
             branch_group = default_ref_options
 
         branches_group = (branch_group, _("Branches"))
         ref_options.append(branches_group)
         choices.extend([x[0] for x in branches_group[0]])
 
         # bookmarks for HG
         if repo.alias == 'hg':
             bookmarks_group = (
                 [(u'book:%s' % safe_unicode(b), safe_unicode(b))
                     for b in repo.bookmarks],
                 _("Bookmarks"))
             ref_options.append(bookmarks_group)
             choices.extend([x[0] for x in bookmarks_group[0]])
 
         # tags
         tags_group = (
             [(u'tag:%s' % safe_unicode(t), safe_unicode(t))
                 for t in repo.tags],
             _("Tags"))
         ref_options.append(tags_group)
         choices.extend([x[0] for x in tags_group[0]])
 
         return choices, ref_options
 
     def get_server_info(self, environ=None):
         server_info = get_system_info(environ)
         return server_info