Switch branches/tags
Nothing to show
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
425 lines (310 sloc) 15.8 KB
# Copyright (C) 2011-2018 ycmd contributors
# This file is part of ycmd.
# ycmd is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
# ycmd is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# GNU General Public License for more details.
# You should have received a copy of the GNU General Public License
# along with ycmd. If not, see <>.
from __future__ import unicode_literals
from __future__ import print_function
from __future__ import division
from __future__ import absolute_import
# Not installing aliases from python-future; it's unreliable and slow.
from builtins import * # noqa
import abc
import threading
from ycmd.completers import completer_utils
from ycmd.responses import NoDiagnosticSupport
from future.utils import with_metaclass
NO_USER_COMMANDS = 'This completer does not define any commands.'
# Number of seconds to block before returning True in PollForMessages
class Completer( with_metaclass( abc.ABCMeta, object ) ):
"""A base class for all Completers in YCM.
Here's several important things you need to know if you're writing a custom
Completer. The following are functions that the Vim part of YCM will be
calling on your Completer:
*Important note about unicode and byte offsets*
Useful background:
Internally, all Python strings are unicode string objects, unless otherwise
converted to 'bytes' using ToBytes. In particular, the line_value and
file_data.contents entries in the request_data are unicode strings.
However, offsets in the API (such as column_num and start_column) are *byte*
offsets into a utf-8 encoded version of the contents of the line or buffer.
Therefore it is *never* safe to perform 'character' arithmetic
(such as '-1' to get the previous 'character') using these byte offsets, and
they cannot *ever* be used to index into line_value or buffer contents
unicode strings.
It is therefore important to ensure that you use the right type of offsets
for the right type of calculation:
- use codepoint offsets and a unicode string for 'character' calculations
- use byte offsets and utf-8 encoded bytes for all other manipulations
ycmd provides the following ways of accessing the source data and offsets:
For working with utf-8 encoded bytes:
- request_data[ 'line_bytes' ] - the line as utf-8 encoded bytes.
- request_data[ 'start_column' ] and request_data[ 'column_num' ].
For working with 'character' manipulations (unicode strings and codepoint
- request_data[ 'line_value' ] - the line as a unicode string.
- request_data[ 'start_codepoint' ] and request_data[ 'column_codepoint' ].
For converting between the two:
- utils.ToBytes
- utils.ByteOffsetToCodepointOffset
- utils.ToUnicode
- utils.CodepointOffsetToByteOffset
Note: The above use of codepoints for 'character' manipulations is not
strictly correct. There are unicode 'characters' which consume multiple
codepoints. However, it is currently considered viable to use a single
codepoint = a single character until such a time as we improve support for
unicode identifiers. The purpose of the above rule is to prevent crashes and
random encoding exceptions, not to fully support unicode identifiers.
*END: Important note about unicode and byte offsets*
ShouldUseNow() is called with the start column of where a potential completion
string should start and the current line (string) the cursor is on. For
instance, if the user's input is '' and the cursor is on the 'r' in
'bar', start_column will be the 1-based byte index of 'b' in the line. Your
implementation of ShouldUseNow() should return True if your semantic completer
should be used and False otherwise.
This is important to get right. You want to return False if you can't provide
completions because then the identifier completer will kick in, and that's
better than nothing.
Note that it's HIGHLY likely that you want to override the ShouldUseNowInner()
function instead of ShouldUseNow() directly (although chances are that you
probably won't have any need to override either). ShouldUseNow() will call
your *Inner version of the function and will also make sure that the
completion cache is taken into account. You'll see this pattern repeated
throughout the Completer API; YCM calls the "main" version of the function and
that function calls the *Inner version while taking into account the cache.
The cache is important and is a nice performance boost. When the user types in
"foo.", your completer will return a list of all member functions and
variables that can be accessed on the "foo" object. The Completer API caches
this list. The user will then continue typing, let's say "". On every
keystroke after the dot, the Completer API will take the cache into account
and will NOT re-query your completer but will in fact provide fuzzy-search on
the candidate strings that were stored in the cache.
ComputeCandidates() is the main entry point when the user types. For
"", the user query is "bar" and completions matching this string should
be shown. It should return the list of candidates. The format of the result
can be a list of strings or a more complicated list of dictionaries. Use
ycmd.responses.BuildCompletionData to build the detailed response. See to see how its used in practice.
Again, you probably want to override ComputeCandidatesInner().
You also need to implement the SupportedFiletypes() function which should
return a list of strings, where the strings are Vim filetypes your completer
supports. is a good example of a "complicated" completer. A good
example of a simple completer is
The On* functions are provided for your convenience. They are called when
their specific events occur. For instance, the identifier completer collects
all the identifiers in the file in OnFileReadyToParse() which gets called when
the user stops typing for 2 seconds (Vim's CursorHold and CursorHoldI events).
One special function is OnUserCommand. It is called when the user uses the
command :YcmCompleter and is passed all extra arguments used on command
invocation (e.g. OnUserCommand(['first argument', 'second'])). This can be
used for completer-specific commands such as reloading external configuration.
Do not override this function. Instead, you need to implement the
GetSubcommandsMap method. It should return a map between the user commands
and the methods of your completer. See the documentation of this method for
more informations on how to implement it.
Override the Shutdown() member function if your Completer subclass needs to do
custom cleanup logic on server shutdown.
If the completer server provides unsolicited messages, such as used in
Language Server Protocol, then you can override the PollForMessagesInner
method. This method is called by the client in the "long poll" fashion to
receive unsolicited messages. The method should block until a message is
available and return a message response when one becomes available, or True if
no message becomes available before the timeout. The return value must be one
of the following:
- a list of messages to send to the client
- True if a timeout occurred, and the poll should be restarted
- False if an error occurred, and no further polling should be attempted
If your completer uses an external server process, then it can be useful to
implement the ServerIsHealthy member function to handle the /healthy request.
This is very useful for the test suite.
If your server is based on the Language Server Protocol (LSP), take a look at
language_server/language_server_completer, which provides most of the work
necessary to get a LSP-based completion engine up and running."""
def __init__( self, user_options ):
self.user_options = user_options
self.min_num_chars = user_options[ 'min_num_of_chars_for_completion' ]
self.max_diagnostics_to_display = user_options[
'max_diagnostics_to_display' ]
self.prepared_triggers = (
user_trigger_map = user_options[ 'semantic_triggers' ],
filetype_set = set( self.SupportedFiletypes() ) )
if user_options[ 'auto_trigger' ] else None )
self._completions_cache = CompletionsCache()
self._max_candidates = user_options[ 'max_num_candidates' ]
# It's highly likely you DON'T want to override this function but the *Inner
# version of it.
def ShouldUseNow( self, request_data ):
if not self.ShouldUseNowInner( request_data ):
return False
# We have to do the cache valid check and get the completions as part of one
# call because we have to ensure a different thread doesn't change the cache
# data.
cache_completions = self._completions_cache.GetCompletionsIfCacheValid(
request_data )
# If None, then the cache isn't valid and we know we should return true
if cache_completions is None:
return True
previous_results_were_valid = bool( cache_completions )
return previous_results_were_valid
def ShouldUseNowInner( self, request_data ):
if not self.prepared_triggers:
return False
current_line = request_data[ 'line_value' ]
start_codepoint = request_data[ 'start_codepoint' ] - 1
column_codepoint = request_data[ 'column_codepoint' ] - 1
filetype = self._CurrentFiletype( request_data[ 'filetypes' ] )
return self.prepared_triggers.MatchesForFiletype(
current_line, start_codepoint, column_codepoint, filetype )
def QueryLengthAboveMinThreshold( self, request_data ):
# Note: calculation in 'characters' not bytes.
query_length = ( request_data[ 'column_codepoint' ] -
request_data[ 'start_codepoint' ] )
return query_length >= self.min_num_chars
# It's highly likely you DON'T want to override this function but the *Inner
# version of it.
def ComputeCandidates( self, request_data ):
if ( not request_data[ 'force_semantic' ] and
not self.ShouldUseNow( request_data ) ):
return []
candidates = self._GetCandidatesFromSubclass( request_data )
return self.FilterAndSortCandidates( candidates, request_data[ 'query' ] )
def _GetCandidatesFromSubclass( self, request_data ):
cache_completions = self._completions_cache.GetCompletionsIfCacheValid(
request_data )
if cache_completions:
return cache_completions
raw_completions = self.ComputeCandidatesInner( request_data )
self._completions_cache.Update( request_data, raw_completions )
return raw_completions
def ComputeCandidatesInner( self, request_data ):
pass # pragma: no cover
def DefinedSubcommands( self ):
subcommands = sorted( self.GetSubcommandsMap().keys() )
# We don't want expose this subcommand because it is not really needed
# for the user but it is useful in tests for tearing down the server
subcommands.remove( 'StopServer' )
except ValueError:
return subcommands
def GetSubcommandsMap( self ):
"""This method should return a dictionary where each key represents the
completer command name and its value is a lambda function of this form:
( self, request_data, args ) -> method
where "method" is the call to the completer method with corresponding
parameters. See the already implemented completers for examples.
- request_data : the request data supplied by the client
- args: any additional command arguments (after the command name). Usually
return {}
def UserCommandsHelpMessage( self ):
subcommands = self.DefinedSubcommands()
if subcommands:
return ( 'Supported commands are:\n' +
'\n'.join( subcommands ) +
'\nSee the docs for information on what they do.' )
return 'This Completer has no supported subcommands.'
def FilterAndSortCandidates( self, candidates, query ):
if not candidates:
return []
# We need to handle both an omni_completer style completer and a server
# style completer
if isinstance( candidates, dict ) and 'words' in candidates:
candidates = candidates[ 'words' ]
sort_property = ''
if isinstance( candidates[ 0 ], dict ):
if 'word' in candidates[ 0 ]:
sort_property = 'word'
elif 'insertion_text' in candidates[ 0 ]:
sort_property = 'insertion_text'
return self.FilterAndSortCandidatesInner( candidates, sort_property, query )
def FilterAndSortCandidatesInner( self, candidates, sort_property, query ):
return completer_utils.FilterAndSortCandidatesWrap(
candidates, sort_property, query, self._max_candidates )
def OnFileReadyToParse( self, request_data ):
pass # pragma: no cover
def OnBufferVisit( self, request_data ):
pass # pragma: no cover
def OnBufferUnload( self, request_data ):
pass # pragma: no cover
def OnInsertLeave( self, request_data ):
pass # pragma: no cover
def OnUserCommand( self, arguments, request_data ):
if not arguments:
raise ValueError( self.UserCommandsHelpMessage() )
command_map = self.GetSubcommandsMap()
command = command_map[ arguments[ 0 ] ]
except KeyError:
raise ValueError( self.UserCommandsHelpMessage() )
return command( self, request_data, arguments[ 1: ] )
def OnCurrentIdentifierFinished( self, request_data ):
pass # pragma: no cover
def GetDiagnosticsForCurrentFile( self, request_data ):
raise NoDiagnosticSupport
def GetDetailedDiagnostic( self, request_data ):
raise NoDiagnosticSupport
def _CurrentFiletype( self, filetypes ):
supported = self.SupportedFiletypes()
for filetype in filetypes:
if filetype in supported:
return filetype
return filetypes[ 0 ]
def SupportedFiletypes( self ):
return set()
def DebugInfo( self, request_data ):
return ''
def Shutdown( self ):
pass # pragma: no cover
def ServerIsReady( self ):
return self.ServerIsHealthy()
def ServerIsHealthy( self ):
"""Called by the /healthy handler to check if the underlying completion
server is started and ready to receive requests. Returns bool."""
return True
def PollForMessages( self, request_data ):
return self.PollForMessagesInner( request_data, MESSAGE_POLL_TIMEOUT )
def PollForMessagesInner( self, request_data, timeout ):
# Most completers don't implement this. It's only required where unsolicited
# messages or diagnostics are supported, such as in the Language Server
# Protocol. As such, the default implementation just returns False, meaning
# that unsolicited messages are not supported for this filetype.
return False
class CompletionsCache( object ):
"""Cache of computed completions for a particular request."""
def __init__( self ):
self._access_lock = threading.Lock()
def Invalidate( self ):
with self._access_lock:
self._request_data = None
self._completions = None
def Update( self, request_data, completions ):
with self._access_lock:
self._request_data = request_data
self._completions = completions
def GetCompletionsIfCacheValid( self, request_data ):
with self._access_lock:
if self._request_data and self._request_data == request_data:
return self._completions
return None