root
/
PyWebIO
mirror of https://github.com/pywebio/PyWebIO.git


			
							123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361
							import os.path
from functools import partial
from contextlib import contextmanager
import ast

import tornado
from tornado import template
from tornado.web import HTTPError, Finish
from tornado.web import StaticFileHandler

from . import page
from .httpbased import HttpHandler
from .tornado import webio_handler, set_ioloop
from .tornado_http import TornadoHttpContext
from .utils import cdn_validation, print_listen_address
from .page import make_applications
from ..session import register_session_implement, CoroutineBasedSession, ThreadBasedSession, Session
from ..utils import get_free_port, STATIC_PATH, parse_file_size

LOCAL_STATIC_URL = '/_pywebio_static'


def filename_ok(f):
    return not f.startswith(('.', '_'))


def identifiers_info(code):
    """Get the identifiers and theirs docstring from python source code.

    :return dict:
    """
    try:
        tree = ast.parse(code)
    except Exception:
        return {}

    if not isinstance(tree, ast.Module):
        return {}

    identifier2doc = {}
    for node in tree.body:
        if isinstance(node, ast.Assign):
            for name in node.targets:
                if hasattr(name, 'id'):
                    identifier2doc[name.id] = ''
        elif isinstance(node, (ast.FunctionDef, ast.AsyncFunctionDef, ast.ClassDef)):
            doc_string = ast.get_docstring(node) or ''
            title = doc_string.split('\n\n')[0]
            identifier2doc[node.name] = title

    return identifier2doc


def valid_and_norm_path(base, subpath):
    """Join the sub-path to base path. This function always ensure the result path is a subpath of base path.

    :param str base: MUST a absolute path
    :param str subpath: sub-path under the `base` path
    :return: normalized result path. None returned if the sub path is not valid
    """
    subpath = subpath.lstrip('/')
    full_path = os.path.normpath(os.path.join(base, subpath))
    if not full_path.startswith(base):
        return None

    parts = subpath.split('/')
    for i in parts:
        if not filename_ok(i):
            return None

    return full_path


_cached_modules = {}


def _get_module(path, reload=False):
    # https://docs.python.org/3/library/importlib.html#importing-a-source-file-directly
    # https://stackoverflow.com/questions/41861427/python-3-5-how-to-dynamically-import-a-module-given-the-full-file-path-in-the
    global _cached_modules
    import importlib.util

    @contextmanager
    def add_to_path(p):
        import sys
        sys.path.append(p)
        try:
            yield
        finally:
            sys.path.remove(p)

    if not reload and path in _cached_modules:
        return _cached_modules[path]

    # import_name will be the `__name__` of the imported module
    import_name = "__pywebio__"

    with add_to_path(os.path.dirname(path)):
        spec = importlib.util.spec_from_file_location(import_name, path, submodule_search_locations=None)
        module = importlib.util.module_from_spec(spec)
        spec.loader.exec_module(module)
    _cached_modules[path] = module
    return module


_app_list_tpl = template.Template("""
<!DOCTYPE html>
<html lang="">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
    <title>{{ title }}</title>
    <meta name="description" content="PyWebIO applications index">
    <style>a{text-decoration:none;display:inline-block;min-width:{{ max_name_width }}ch}span{color:grey}</style>
</head>
<body>
<h1>{{ title }}</h1>
<hr>
<pre style="line-height: 1.6em; font-size: 16px;">
{% for name,doc in files %} <a href="{{ name }}">{{ name }}</a>  <span>{{ doc }}</span> 
{% end %}</pre>
<hr>
</body>
</html>
""".strip())


def default_index_page(path, base):
    urlpath = path[len(base):] or '/'
    title = "Index of %s" % urlpath
    dirs = [] if path == base else [('../', '')]  # (name, doc)
    files = []  # (name, doc)
    for f in os.listdir(path):
        if not filename_ok(f):
            continue

        full_path = os.path.join(path, f)
        if os.path.isfile(full_path):
            if f.endswith('.py'):
                code = open(full_path, encoding='utf8').read()
                identifiers = identifiers_info(code)
                if 'main' in identifiers:
                    files.append([f[:-3], identifiers['main']])
        else:
            dirs.append([(f + '/'), ''])

    items = dirs + files
    max_name_width = max([len(n) for n, _ in items]+[0])
    return _app_list_tpl.generate(files=items, title=title, max_name_width=max_name_width)


def get_app_from_path(request_path, base, index, reload=False):
    """Get PyWebIO app

    :param str request_path: request path
    :param str base: dir base path, MUST a absolute path
    :param callable index:
    :return: ('error', http error code in int) / ('app', pywebio task function) / ('html', Html content in bytes)
    """
    path = valid_and_norm_path(base, request_path)
    if path is None:
        return 'error', 403

    if os.path.isdir(path):
        if not request_path.endswith('/'):
            return 'error', 404

        if os.path.isfile(os.path.join(path, 'index.py')):
            path = os.path.join(path, 'index.py')
        elif index:
            content = index(path)
            return 'html', content
        else:
            return 'error', 404
    else:
        path += '.py'

    if not os.path.isfile(path):
        return 'error', 404

    module = _get_module(path, reload=reload)
    if hasattr(module, 'main'):
        return 'app', make_applications(module.main)

    return 'error', 404


def _path_deploy(base, port=0, host='', static_dir=None, max_payload_size=2 ** 20 * 200,
                 **tornado_app_settings):
    if not host:
        host = '0.0.0.0'

    if port == 0:
        port = get_free_port()

    tornado_app_settings = {k: v for k, v in tornado_app_settings.items() if v is not None}

    abs_base = os.path.normpath(os.path.abspath(base))

    register_session_implement(CoroutineBasedSession)
    register_session_implement(ThreadBasedSession)

    RequestHandler = yield abs_base

    handlers = []
    if static_dir is not None:
        handlers.append((r"/static/(.*)", StaticFileHandler, {"path": static_dir}))
    handlers.append((LOCAL_STATIC_URL+r"/(.*)", StaticFileHandler, {"path": STATIC_PATH}))
    handlers.append((r"/.*", RequestHandler))

    print_listen_address(host, port)

    set_ioloop(tornado.ioloop.IOLoop.current())  # to enable bokeh app
    app = tornado.web.Application(handlers=handlers, **tornado_app_settings)
    app.listen(port, address=host, max_buffer_size=max_payload_size)
    tornado.ioloop.IOLoop.current().start()


def path_deploy(base, port=0, host='',
                index=True, static_dir=None,
                reconnect_timeout=0,
                cdn=True, debug=False,
                allowed_origins=None, check_origin=None,
                max_payload_size='200M',
                **tornado_app_settings):
    """Deploy the PyWebIO applications from a directory.

    The server communicates with the browser using WebSocket protocol.

    :param str base: Base directory to load PyWebIO application.
    :param int port: The port the server listens on.
    :param str host: The host the server listens on.
    :param bool/callable index: Whether to provide a default index page when request a directory, default is ``True``.
       ``index`` also accepts a function to custom index page, which receives the requested directory path as parameter
       and return HTML content in string.

       You can override the index page by add a `index.py` PyWebIO app file to the directory.
    :param str static_dir: Directory to store the application static files.
       The files in this directory can be accessed via ``http://<host>:<port>/static/files``.
       For example, if there is a ``A/B.jpg`` file in ``static_dir`` path,
       it can be accessed via ``http://<host>:<port>/static/A/B.jpg``.
    :param int reconnect_timeout: The client can reconnect to server within ``reconnect_timeout`` seconds after an unexpected disconnection.
       If set to 0 (default), once the client disconnects, the server session will be closed.

    The rest arguments of ``path_deploy()`` have the same meaning as for :func:`pywebio.platform.tornado.start_server`
    """
    debug = Session.debug = os.environ.get('PYWEBIO_DEBUG', debug)
    page.MAX_PAYLOAD_SIZE = max_payload_size = parse_file_size(max_payload_size)
    # Since some cloud server may close idle connections (such as heroku),
    # use `websocket_ping_interval` to  keep the connection alive
    tornado_app_settings.setdefault('websocket_ping_interval', 30)
    tornado_app_settings.setdefault('websocket_max_message_size', max_payload_size)  # Backward compatible
    tornado_app_settings['websocket_max_message_size'] = parse_file_size(tornado_app_settings['websocket_max_message_size'])
    gen = _path_deploy(base, port=port, host=host,
                       static_dir=static_dir, debug=debug,
                       max_payload_size=max_payload_size,
                       **tornado_app_settings)

    cdn = cdn_validation(cdn, 'warn', stacklevel=3)  # if CDN is not available, warn user and disable CDN

    abs_base = next(gen)

    index_func = {True: partial(default_index_page, base=abs_base), False: lambda p: '403 Forbidden'}.get(index, index)

    Handler = webio_handler(lambda: None, cdn=cdn, allowed_origins=allowed_origins,
                            check_origin=check_origin, reconnect_timeout=reconnect_timeout)

    class WSHandler(Handler):

        def get_cdn(self):
            _cdn = super().get_cdn()
            if not _cdn:
                return LOCAL_STATIC_URL
            return _cdn

        def get_app(self):
            reload = self.get_query_argument('reload', None) is not None
            type, res = get_app_from_path(self.request.path, abs_base, index=index_func, reload=reload)
            if type == 'error':
                raise HTTPError(status_code=res)
            elif type == 'html':
                raise Finish(res)

            app_name = self.get_query_argument('app', 'index')
            app = res.get(app_name) or res['index']

            return app

    gen.send(WSHandler)
    gen.close()


def path_deploy_http(base, port=0, host='',
                     index=True, static_dir=None,
                     cdn=True, debug=False,
                     allowed_origins=None, check_origin=None,
                     session_expire_seconds=None,
                     session_cleanup_interval=None,
                     max_payload_size='200M',
                     **tornado_app_settings):
    """Deploy the PyWebIO applications from a directory.

    The server communicates with the browser using HTTP protocol.

    The ``base``, ``port``, ``host``, ``index``, ``static_dir`` arguments of ``path_deploy_http()``
    have the same meaning as for :func:`pywebio.platform.path_deploy`

    The rest arguments of ``path_deploy_http()`` have the same meaning as for :func:`pywebio.platform.tornado_http.start_server`
    """
    debug = Session.debug = os.environ.get('PYWEBIO_DEBUG', debug)

    page.MAX_PAYLOAD_SIZE = max_payload_size = parse_file_size(max_payload_size)

    gen = _path_deploy(base, port=port, host=host,
                       static_dir=static_dir, debug=debug,
                       max_payload_size=max_payload_size,
                       **tornado_app_settings)

    cdn = cdn_validation(cdn, 'warn', stacklevel=3)  # if CDN is not available, warn user and disable CDN

    abs_base = next(gen)
    index_func = {True: partial(default_index_page, base=abs_base), False: lambda p: '403 Forbidden'}.get(index, index)

    def get_app(context: TornadoHttpContext):
        reload = context.request_url_parameter('reload', None) is not None

        type, res = get_app_from_path(context.get_path(), abs_base, index=index_func, reload=reload)
        if type == 'error':
            raise HTTPError(status_code=res)
        elif type == 'html':
            raise Finish(res)

        app_name = context.request_url_parameter('app', 'index')
        return res.get(app_name) or res['index']

    class HttpPathDeployHandler(HttpHandler):
        def get_cdn(self, context):
            _cdn = super().get_cdn(context)
            if not _cdn:
                return LOCAL_STATIC_URL
            return _cdn

    handler = HttpPathDeployHandler(app_loader=get_app, cdn=cdn,
                                    session_expire_seconds=session_expire_seconds,
                                    session_cleanup_interval=session_cleanup_interval,
                                    allowed_origins=allowed_origins,
                                    check_origin=check_origin)

    class ReqHandler(tornado.web.RequestHandler):
        def options(self):
            return self.get()

        def post(self):
            return self.get()

        def get(self):
            context = TornadoHttpContext(self)
            self.write(handler.handle_request(context))

    gen.send(ReqHandler)
    gen.close()