# Copyright 2017 The TensorFlow Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
"""TensorBoard WSGI Application Logic.
Provides TensorBoardWSGIApp for building a TensorBoard WSGI app.
"""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import atexit
import base64
import collections
import contextlib
import hashlib
import json
import os
import re
import shutil
import tempfile
import textwrap
import threading
import time
import six
from six.moves.urllib import (
parse as urlparse,
) # pylint: disable=wrong-import-order
from werkzeug import wrappers
from tensorboard import errors
from tensorboard import plugin_util
from tensorboard.backend import empty_path_redirect
from tensorboard.backend import experiment_id
from tensorboard.backend import experimental_plugin
from tensorboard.backend import http_util
from tensorboard.backend import path_prefix
from tensorboard.backend import security_validator
from tensorboard.backend.event_processing import (
data_provider as event_data_provider,
)
from tensorboard.backend.event_processing import (
plugin_event_multiplexer as event_multiplexer,
)
from tensorboard.backend.event_processing import tag_types
from tensorboard.plugins import base_plugin
from tensorboard.plugins.audio import metadata as audio_metadata
from tensorboard.plugins.core import core_plugin
from tensorboard.plugins.histogram import metadata as histogram_metadata
from tensorboard.plugins.image import metadata as image_metadata
from tensorboard.plugins.pr_curve import metadata as pr_curve_metadata
from tensorboard.plugins.scalar import metadata as scalar_metadata
from tensorboard.util import tb_logging
DEFAULT_SIZE_GUIDANCE = {
tag_types.TENSORS: 10,
}
# TODO(@wchargin): Once SQL mode is in play, replace this with an
# alternative that does not privilege first-party plugins.
DEFAULT_TENSOR_SIZE_GUIDANCE = {
scalar_metadata.PLUGIN_NAME: 1000,
image_metadata.PLUGIN_NAME: 10,
audio_metadata.PLUGIN_NAME: 10,
histogram_metadata.PLUGIN_NAME: 500,
pr_curve_metadata.PLUGIN_NAME: 100,
}
DATA_PREFIX = "/data"
PLUGIN_PREFIX = "/plugin"
PLUGINS_LISTING_ROUTE = "/plugins_listing"
PLUGIN_ENTRY_ROUTE = "/plugin_entry.html"
EXPERIMENTAL_PLUGINS_QUERY_PARAM = "experimentalPlugin"
# Slashes in a plugin name could throw the router for a loop. An empty
# name would be confusing, too. To be safe, let's restrict the valid
# names as follows.
_VALID_PLUGIN_RE = re.compile(r"^[A-Za-z0-9_-]+$")
logger = tb_logging.get_logger()
def _apply_tensor_size_guidance(sampling_hints):
"""Apply user per-summary size guidance overrides."""
tensor_size_guidance = dict(DEFAULT_TENSOR_SIZE_GUIDANCE)
tensor_size_guidance.update(sampling_hints)
return tensor_size_guidance
def standard_tensorboard_wsgi(flags, plugin_loaders, assets_zip_provider):
"""Construct a TensorBoardWSGIApp with standard plugins and multiplexer.
Args:
flags: An argparse.Namespace containing TensorBoard CLI flags.
plugin_loaders: A list of TBLoader instances.
assets_zip_provider: See TBContext documentation for more information.
Returns:
The new TensorBoard WSGI application.
:type plugin_loaders: list[base_plugin.TBLoader]
:rtype: TensorBoardWSGI
"""
data_provider = None
multiplexer = None
reload_interval = flags.reload_interval
# Regular logdir loading mode.
sampling_hints = flags.samples_per_plugin
multiplexer = event_multiplexer.EventMultiplexer(
size_guidance=DEFAULT_SIZE_GUIDANCE,
tensor_size_guidance=_apply_tensor_size_guidance(sampling_hints),
purge_orphaned_data=flags.purge_orphaned_data,
max_reload_threads=flags.max_reload_threads,
event_file_active_filter=_get_event_file_active_filter(flags),
)
data_provider = event_data_provider.MultiplexerDataProvider(
multiplexer, flags.logdir or flags.logdir_spec
)
if reload_interval >= 0:
# We either reload the multiplexer once when TensorBoard starts up, or we
# continuously reload the multiplexer.
if flags.logdir:
path_to_run = {os.path.expanduser(flags.logdir): None}
else:
path_to_run = parse_event_files_spec(flags.logdir_spec)
start_reloading_multiplexer(
multiplexer, path_to_run, reload_interval, flags.reload_task
)
return TensorBoardWSGIApp(
flags, plugin_loaders, data_provider, assets_zip_provider, multiplexer
)
def _handling_errors(wsgi_app):
def wrapper(*args):
(environ, start_response) = (args[-2], args[-1])
try:
return wsgi_app(*args)
except errors.PublicError as e:
request = wrappers.Request(environ)
error_app = http_util.Respond(
request, str(e), "text/plain", code=e.http_code
)
return error_app(environ, start_response)
# Let other exceptions be handled by the server, as an opaque
# internal server error.
return wrapper
def TensorBoardWSGIApp(
flags,
plugins,
data_provider=None,
assets_zip_provider=None,
deprecated_multiplexer=None,
):
"""Constructs a TensorBoard WSGI app from plugins and data providers.
Args:
flags: An argparse.Namespace containing TensorBoard CLI flags.
plugins: A list of plugin loader instances.
assets_zip_provider: See TBContext documentation for more information.
data_provider: Instance of `tensorboard.data.provider.DataProvider`. May
be `None` if `flags.generic_data` is set to `"false"` in which case
`deprecated_multiplexer` must be passed instead.
deprecated_multiplexer: Optional `plugin_event_multiplexer.EventMultiplexer`
to use for any plugins not yet enabled for the DataProvider API.
Required if the data_provider argument is not passed.
Returns:
A WSGI application that implements the TensorBoard backend.
:type plugins: list[base_plugin.TBLoader]
"""
plugin_name_to_instance = {}
context = base_plugin.TBContext(
data_provider=data_provider,
flags=flags,
logdir=flags.logdir,
multiplexer=deprecated_multiplexer,
assets_zip_provider=assets_zip_provider,
plugin_name_to_instance=plugin_name_to_instance,
sampling_hints=flags.samples_per_plugin,
window_title=flags.window_title,
)
tbplugins = []
experimental_plugins = []
for plugin_spec in plugins:
loader = make_plugin_loader(plugin_spec)
try:
plugin = loader.load(context)
except Exception:
logger.error(
"Failed to load plugin %s; ignoring it.",
getattr(loader.load, "__qualname__", loader.load),
exc_info=True,
)
plugin = None
if plugin is None:
continue
tbplugins.append(plugin)
if isinstance(
loader, experimental_plugin.ExperimentalPlugin
) or isinstance(plugin, experimental_plugin.ExperimentalPlugin):
experimental_plugins.append(plugin.plugin_name)
plugin_name_to_instance[plugin.plugin_name] = plugin
return TensorBoardWSGI(
tbplugins, flags.path_prefix, data_provider, experimental_plugins
)
class TensorBoardWSGI(object):
"""The TensorBoard WSGI app that delegates to a set of TBPlugin."""
def __init__(
self,
plugins,
path_prefix="",
data_provider=None,
experimental_plugins=None,
):
"""Constructs TensorBoardWSGI instance.
Args:
plugins: A list of base_plugin.TBPlugin subclass instances.
flags: An argparse.Namespace containing TensorBoard CLI flags.
data_provider: `tensorboard.data.provider.DataProvider` or
`None`; if present, will inform the "active" state of
`/plugins_listing`.
experimental_plugins: A list of plugin names that are only provided
experimentally. The corresponding plugins will only be activated for
a user if the user has specified the plugin with the experimentalPlugin
query parameter in the URL.
Returns:
A WSGI application for the set of all TBPlugin instances.
Raises:
ValueError: If some plugin has no plugin_name
ValueError: If some plugin has an invalid plugin_name (plugin
names must only contain [A-Za-z0-9_.-])
ValueError: If two plugins have the same plugin_name
ValueError: If some plugin handles a route that does not start
with a slash
:type plugins: list[base_plugin.TBPlugin]
"""
self._plugins = plugins
self._path_prefix = path_prefix
self._data_provider = data_provider
self._experimental_plugins = frozenset(experimental_plugins or ())
if self._path_prefix.endswith("/"):
# Should have been fixed by `fix_flags`.
raise ValueError(
"Trailing slash in path prefix: %r" % self._path_prefix
)
self.exact_routes = {
# TODO(@chihuahua): Delete this RPC once we have skylark rules that
# obviate the need for the frontend to determine which plugins are
# active.
DATA_PREFIX + PLUGINS_LISTING_ROUTE: self._serve_plugins_listing,
DATA_PREFIX + PLUGIN_ENTRY_ROUTE: self._serve_plugin_entry,
}
unordered_prefix_routes = {}
# Serve the routes from the registered plugins using their name as the route
# prefix. For example if plugin z has two routes /a and /b, they will be
# served as /data/plugin/z/a and /data/plugin/z/b.
plugin_names_encountered = set()
for plugin in self._plugins:
if plugin.plugin_name is None:
raise ValueError("Plugin %s has no plugin_name" % plugin)
if not _VALID_PLUGIN_RE.match(plugin.plugin_name):
raise ValueError(
"Plugin %s has invalid name %r"
% (plugin, plugin.plugin_name)
)
if plugin.plugin_name in plugin_names_encountered:
raise ValueError(
"Duplicate plugins for name %s" % plugin.plugin_name
)
plugin_names_encountered.add(plugin.plugin_name)
try:
plugin_apps = plugin.get_plugin_apps()
except Exception as e: # pylint: disable=broad-except
if (
type(plugin) is core_plugin.CorePlugin
): # pylint: disable=unidiomatic-typecheck
raise
logger.warn(
"Plugin %s failed. Exception: %s",
plugin.plugin_name,
str(e),
)
continue
for route, app in plugin_apps.items():
if not route.startswith("/"):
raise ValueError(
"Plugin named %r handles invalid route %r: "
"route does not start with a slash"
% (plugin.plugin_name, route)
)
if (
type(plugin) is core_plugin.CorePlugin
): # pylint: disable=unidiomatic-typecheck
path = route
else:
path = (
DATA_PREFIX
+ PLUGIN_PREFIX
+ "/"
+ plugin.plugin_name
+ route
)
if path.endswith("/*"):
# Note we remove the '*' but leave the slash in place.
path = path[:-1]
if "*" in path:
# note we re-add the removed * in the format string
raise ValueError(
"Plugin %r handles invalid route '%s*': Only "
"trailing wildcards are supported "
"(i.e., `/.../*`)" % (plugin.plugin_name, path)
)
unordered_prefix_routes[path] = app
else:
if "*" in path:
raise ValueError(
"Plugin %r handles invalid route %r: Only "
"trailing wildcards are supported "
"(i.e., `/.../*`)" % (plugin.plugin_name, path)
)
self.exact_routes[path] = app
# Wildcard routes will be checked in the given order, so we sort them
# longest to shortest so that a more specific route will take precedence
# over a more general one (e.g., a catchall route `/*` should come last).
self.prefix_routes = collections.OrderedDict(
sorted(
six.iteritems(unordered_prefix_routes),
key=lambda x: len(x[0]),
reverse=True,
)
)
self._app = self._create_wsgi_app()
def _create_wsgi_app(self):
"""Apply middleware to create the final WSGI app."""
app = self._route_request
app = empty_path_redirect.EmptyPathRedirectMiddleware(app)
app = experiment_id.ExperimentIdMiddleware(app)
app = path_prefix.PathPrefixMiddleware(app, self._path_prefix)
app = security_validator.SecurityValidatorMiddleware(app)
app = _handling_errors(app)
return app
@wrappers.Request.application
def _serve_plugin_entry(self, request):
"""Serves a HTML for iframed plugin entry point.
Args:
request: The werkzeug.Request object.
Returns:
A werkzeug.Response object.
"""
name = request.args.get("name")
plugins = [
plugin for plugin in self._plugins if plugin.plugin_name == name
]
if not plugins:
raise errors.NotFoundError(name)
if len(plugins) > 1:
# Technically is not possible as plugin names are unique and is checked
# by the check on __init__.
reason = (
"Plugin invariant error: multiple plugins with name "
"{name} found: {list}"
).format(name=name, list=plugins)
raise AssertionError(reason)
plugin = plugins[0]
module_path = plugin.frontend_metadata().es_module_path
if not module_path:
return http_util.Respond(
request, "Plugin is not module loadable", "text/plain", code=400
)
# non-self origin is blocked by CSP but this is a good invariant checking.
if urlparse.urlparse(module_path).netloc:
raise ValueError("Expected es_module_path to be non-absolute path")
module_json = json.dumps("." + module_path)
script_content = "import({}).then((m) => void m.render());".format(
module_json
)
digest = hashlib.sha256(script_content.encode("utf-8")).digest()
script_sha = base64.b64encode(digest).decode("ascii")
html = textwrap.dedent(
"""
"""
).format(name=name, script_content=script_content)
return http_util.Respond(
request, html, "text/html", csp_scripts_sha256s=[script_sha],
)
@wrappers.Request.application
def _serve_plugins_listing(self, request):
"""Serves an object mapping plugin name to whether it is enabled.
Args:
request: The werkzeug.Request object.
Returns:
A werkzeug.Response object.
"""
response = collections.OrderedDict()
eid = plugin_util.experiment_id(request.environ)
plugins_with_data = frozenset(
self._data_provider.list_plugins(eid) or frozenset()
if self._data_provider is not None
else frozenset()
)
plugins_to_skip = self._experimental_plugins - frozenset(
request.args.getlist(EXPERIMENTAL_PLUGINS_QUERY_PARAM)
)
for plugin in self._plugins:
if plugin.plugin_name in plugins_to_skip:
continue
if (
type(plugin) is core_plugin.CorePlugin
): # pylint: disable=unidiomatic-typecheck
# This plugin's existence is a backend implementation detail.
continue
is_active = bool(
frozenset(plugin.data_plugin_names()) & plugins_with_data
)
if not is_active:
try:
start = time.time()
is_active = plugin.is_active()
elapsed = time.time() - start
logger.info(
"Plugin listing: is_active() for %s took %0.3f seconds",
plugin.plugin_name,
elapsed,
)
except Exception:
is_active = False
logger.error(
"Plugin listing: is_active() for %s failed (marking inactive)",
plugin.plugin_name,
exc_info=True,
)
plugin_metadata = plugin.frontend_metadata()
output_metadata = {
"disable_reload": plugin_metadata.disable_reload,
"enabled": is_active,
# loading_mechanism set below
"remove_dom": plugin_metadata.remove_dom,
# tab_name set below
}
if plugin_metadata.tab_name is not None:
output_metadata["tab_name"] = plugin_metadata.tab_name
else:
output_metadata["tab_name"] = plugin.plugin_name
es_module_handler = plugin_metadata.es_module_path
element_name = plugin_metadata.element_name
is_ng_component = plugin_metadata.is_ng_component
if is_ng_component:
if element_name is not None:
raise ValueError(
"Plugin %r declared as both Angular built-in and legacy"
% plugin.plugin_name
)
if es_module_handler is not None:
raise ValueError(
"Plugin %r declared as both Angular built-in and iframed"
% plugin.plugin_name
)
loading_mechanism = {
"type": "NG_COMPONENT",
}
elif element_name is not None and es_module_handler is not None:
logger.error(
"Plugin %r declared as both legacy and iframed; skipping",
plugin.plugin_name,
)
continue
elif element_name is not None and es_module_handler is None:
loading_mechanism = {
"type": "CUSTOM_ELEMENT",
"element_name": element_name,
}
elif element_name is None and es_module_handler is not None:
loading_mechanism = {
"type": "IFRAME",
"module_path": "".join(
[
request.script_root,
DATA_PREFIX,
PLUGIN_PREFIX,
"/",
plugin.plugin_name,
es_module_handler,
]
),
}
else:
# As a compatibility measure (for plugins that we don't
# control), we'll pull it from the frontend registry for now.
loading_mechanism = {
"type": "NONE",
}
output_metadata["loading_mechanism"] = loading_mechanism
response[plugin.plugin_name] = output_metadata
return http_util.Respond(request, response, "application/json")
def __call__(self, environ, start_response):
"""Central entry point for the TensorBoard application.
This __call__ method conforms to the WSGI spec, so that instances of this
class are WSGI applications.
Args:
environ: See WSGI spec (PEP 3333).
start_response: See WSGI spec (PEP 3333).
"""
return self._app(environ, start_response)
def _route_request(self, environ, start_response):
"""Delegate an incoming request to sub-applications.
This method supports strict string matching and wildcard routes of a
single path component, such as `/foo/*`. Other routing patterns,
like regular expressions, are not supported.
This is the main TensorBoard entry point before middleware is
applied. (See `_create_wsgi_app`.)
Args:
environ: See WSGI spec (PEP 3333).
start_response: See WSGI spec (PEP 3333).
"""
request = wrappers.Request(environ)
parsed_url = urlparse.urlparse(request.path)
clean_path = _clean_path(parsed_url.path)
# pylint: disable=too-many-function-args
if clean_path in self.exact_routes:
return self.exact_routes[clean_path](environ, start_response)
else:
for path_prefix in self.prefix_routes:
if clean_path.startswith(path_prefix):
return self.prefix_routes[path_prefix](
environ, start_response
)
logger.warn("path %s not found, sending 404", clean_path)
return http_util.Respond(
request, "Not found", "text/plain", code=404
)(environ, start_response)
# pylint: enable=too-many-function-args
def parse_event_files_spec(logdir_spec):
"""Parses `logdir_spec` into a map from paths to run group names.
The `--logdir_spec` flag format is a comma-separated list of path
specifications. A path spec looks like 'group_name:/path/to/directory' or
'/path/to/directory'; in the latter case, the group is unnamed. Group names
cannot start with a forward slash: /foo:bar/baz will be interpreted as a spec
with no name and path '/foo:bar/baz'.
Globs are not supported.
Args:
logdir: A comma-separated list of run specifications.
Returns:
A dict mapping directory paths to names like {'/path/to/directory': 'name'}.
Groups without an explicit name are named after their path. If logdir is
None, returns an empty dict, which is helpful for testing things that don't
require any valid runs.
"""
files = {}
if logdir_spec is None:
return files
# Make sure keeping consistent with ParseURI in core/lib/io/path.cc
uri_pattern = re.compile("[a-zA-Z][0-9a-zA-Z.]*://.*")
for specification in logdir_spec.split(","):
# Check if the spec contains group. A spec start with xyz:// is regarded as
# URI path spec instead of group spec. If the spec looks like /foo:bar/baz,
# then we assume it's a path with a colon. If the spec looks like
# [a-zA-z]:\foo then we assume its a Windows path and not a single letter
# group
if (
uri_pattern.match(specification) is None
and ":" in specification
and specification[0] != "/"
and not os.path.splitdrive(specification)[0]
):
# We split at most once so run_name:/path:with/a/colon will work.
run_name, _, path = specification.partition(":")
else:
run_name = None
path = specification
if uri_pattern.match(path) is None:
path = os.path.realpath(os.path.expanduser(path))
files[path] = run_name
return files
def start_reloading_multiplexer(
multiplexer, path_to_run, load_interval, reload_task
):
"""Starts automatically reloading the given multiplexer.
If `load_interval` is positive, the thread will reload the multiplexer
by calling `ReloadMultiplexer` every `load_interval` seconds, starting
immediately. Otherwise, reloads the multiplexer once and never again.
Args:
multiplexer: The `EventMultiplexer` to add runs to and reload.
path_to_run: A dict mapping from paths to run names, where `None` as the run
name is interpreted as a run name equal to the path.
load_interval: An integer greater than or equal to 0. If positive, how many
seconds to wait after one load before starting the next load. Otherwise,
reloads the multiplexer once and never again (no continuous reloading).
reload_task: Indicates the type of background task to reload with.
Raises:
ValueError: If `load_interval` is negative.
"""
if load_interval < 0:
raise ValueError("load_interval is negative: %d" % load_interval)
def _reload():
while True:
start = time.time()
logger.info("TensorBoard reload process beginning")
for path, name in six.iteritems(path_to_run):
multiplexer.AddRunsFromDirectory(path, name)
logger.info(
"TensorBoard reload process: Reload the whole Multiplexer"
)
multiplexer.Reload()
duration = time.time() - start
logger.info(
"TensorBoard done reloading. Load took %0.3f secs", duration
)
if load_interval == 0:
# Only load the multiplexer once. Do not continuously reload.
break
time.sleep(load_interval)
if reload_task == "process":
logger.info("Launching reload in a child process")
import multiprocessing
process = multiprocessing.Process(target=_reload, name="Reloader")
# Best-effort cleanup; on exit, the main TB parent process will attempt to
# kill all its daemonic children.
process.daemon = True
process.start()
elif reload_task in ("thread", "auto"):
logger.info("Launching reload in a daemon thread")
thread = threading.Thread(target=_reload, name="Reloader")
# Make this a daemon thread, which won't block TB from exiting.
thread.daemon = True
thread.start()
elif reload_task == "blocking":
if load_interval != 0:
raise ValueError(
"blocking reload only allowed with load_interval=0"
)
_reload()
else:
raise ValueError("unrecognized reload_task: %s" % reload_task)
def _clean_path(path):
"""Removes a trailing slash from a non-root path.
Arguments:
path: The path of a request.
Returns:
The route to use to serve the request.
"""
if path != "/" and path.endswith("/"):
return path[:-1]
return path
def _get_event_file_active_filter(flags):
"""Returns a predicate for whether an event file load timestamp is active.
Returns:
A predicate function accepting a single UNIX timestamp float argument, or
None if multi-file loading is not enabled.
"""
if not flags.reload_multifile:
return None
inactive_secs = flags.reload_multifile_inactive_secs
if inactive_secs == 0:
return None
if inactive_secs < 0:
return lambda timestamp: True
return lambda timestamp: timestamp + inactive_secs >= time.time()
def make_plugin_loader(plugin_spec):
"""Returns a plugin loader for the given plugin.
Args:
plugin_spec: A TBPlugin subclass, or a TBLoader instance or subclass.
Returns:
A TBLoader for the given plugin.
:type plugin_spec:
Type[base_plugin.TBPlugin] | Type[base_plugin.TBLoader] |
base_plugin.TBLoader
:rtype: base_plugin.TBLoader
"""
if isinstance(plugin_spec, base_plugin.TBLoader):
return plugin_spec
if isinstance(plugin_spec, type):
if issubclass(plugin_spec, base_plugin.TBLoader):
return plugin_spec()
if issubclass(plugin_spec, base_plugin.TBPlugin):
return base_plugin.BasicLoader(plugin_spec)
raise TypeError("Not a TBLoader or TBPlugin subclass: %r" % (plugin_spec,))