summaryrefslogtreecommitdiffstats
path: root/utils/test/testapi/opnfv_testapi/tornado_swagger/swagger.py
blob: 3d21edefb053bb0d5587954143e7a3c01144b50d (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
##############################################################################
# Copyright (c) 2016 ZTE Corporation
# feng.xiaowei@zte.com.cn
# All rights reserved. This program and the accompanying materials
# are made available under the terms of the Apache License, Version 2.0
# which accompanies this distribution, and is available at
# http://www.apache.org/licenses/LICENSE-2.0
##############################################################################
import inspect
from functools import wraps
from HTMLParser import HTMLParser

import epydoc.markup
import tornado.web

from settings import default_settings, models
from handlers import swagger_handlers


class EpytextParser(HTMLParser):
    a_text = False

    def __init__(self, tag):
        HTMLParser.__init__(self)
        self.tag = tag
        self.data = None

    def handle_starttag(self, tag, attr):
        if tag == self.tag:
            self.a_text = True

    def handle_endtag(self, tag):
        if tag == self.tag:
            self.a_text = False

    def handle_data(self, data):
        if self.a_text:
            self.data = data

    def get_data(self):
        return self.data


class DocParser(object):
    def __init__(self):
        self.notes = None
        self.summary = None
        self.responseClass = None
        self.responseMessages = []
        self.params = {}
        self.properties = {}

    def parse_docstring(self, text):
        if text is None:
            return

        errors = []
        doc = epydoc.markup.parse(text, markup='epytext', errors=errors)
        _, fields = doc.split_fields(errors)

        for field in fields:
            tag = field.tag()
            arg = field.arg()
            body = field.body()
            self._get_parser(tag)(arg=arg, body=body)
        return doc

    def _get_parser(self, tag):
        parser = {
            'param': self._parse_param,
            'type': self._parse_type,
            'in': self._parse_in,
            'required': self._parse_required,
            'rtype': self._parse_rtype,
            'property': self._parse_property,
            'ptype': self._parse_ptype,
            'return': self._parse_return,
            'raise': self._parse_return,
            'notes': self._parse_notes,
            'description': self._parse_description,
        }
        return parser.get(tag, self._not_supported)

    def _parse_param(self, **kwargs):
        arg = kwargs.get('arg', None)
        body = self._get_body(**kwargs)
        self.params.setdefault(arg, {}).update({
            'name': arg,
            'description': body,
        })

        if 'paramType' not in self.params[arg]:
            self.params[arg]['paramType'] = 'query'

    def _parse_type(self, **kwargs):
        arg = kwargs.get('arg', None)
        body = self._get_body(**kwargs)
        self.params.setdefault(arg, {}).update({
            'name': arg,
            'dataType': body
        })

    def _parse_in(self, **kwargs):
        arg = kwargs.get('arg', None)
        body = self._get_body(**kwargs)
        self.params.setdefault(arg, {}).update({
            'name': arg,
            'paramType': body
        })

    def _parse_required(self, **kwargs):
        arg = kwargs.get('arg', None)
        body = self._get_body(**kwargs)
        self.params.setdefault(arg, {}).update({
            'name': arg,
            'required': False if body in ['False', 'false'] else True
        })

    def _parse_rtype(self, **kwargs):
        body = self._get_body(**kwargs)
        self.responseClass = body

    def _parse_property(self, **kwargs):
        arg = kwargs.get('arg', None)
        self.properties.setdefault(arg, {}).update({
            'type': 'string'
        })

    def _parse_ptype(self, **kwargs):
        arg = kwargs.get('arg', None)
        code = self._parse_epytext_para('code', **kwargs)
        link = self._parse_epytext_para('link', **kwargs)
        if code is None:
            self.properties.setdefault(arg, {}).update({
                'type': link
            })
        elif code == 'list':
            self.properties.setdefault(arg, {}).update({
                'type': 'array',
                'items': {'type': link}
            })

    def _parse_return(self, **kwargs):
        arg = kwargs.get('arg', None)
        body = self._get_body(**kwargs)
        self.responseMessages.append({
            'code': arg,
            'message': body
        })

    def _parse_notes(self, **kwargs):
        body = self._get_body(**kwargs)
        self.notes = self._sanitize_doc(body)

    def _parse_description(self, **kwargs):
        body = self._get_body(**kwargs)
        self.summary = self._sanitize_doc(body)

    def _not_supported(self, **kwargs):
        pass

    @staticmethod
    def _sanitize_doc(comment):
        return comment.replace('\n', '<br/>') if comment else comment

    @staticmethod
    def _get_body(**kwargs):
        body = kwargs.get('body', None)
        return body.to_plaintext(None).strip() if body else body

    @staticmethod
    def _parse_epytext_para(tag, **kwargs):
        def _parse_epytext(tag, body):
            epytextParser = EpytextParser(tag)
            epytextParser.feed(str(body))
            data = epytextParser.get_data()
            epytextParser.close()
            return data

        body = kwargs.get('body', None)
        return _parse_epytext(tag, body) if body else body


class model(DocParser):
    def __init__(self, *args, **kwargs):
        super(model, self).__init__()
        self.args = args
        self.kwargs = kwargs
        self.required = []
        self.cls = None

    def __call__(self, *args):
        if self.cls:
            return self.cls

        cls = args[0]
        self._parse_model(cls)

        return cls

    def _parse_model(self, cls):
        self.id = cls.__name__
        self.cls = cls
        if '__init__' in dir(cls):
            self._parse_args(cls.__init__)
        self.parse_docstring(inspect.getdoc(cls))
        models.append(self)

    def _parse_args(self, func):
        argspec = inspect.getargspec(func)
        argspec.args.remove("self")
        defaults = {}
        if argspec.defaults:
            defaults = list(zip(argspec.args[-len(argspec.defaults):],
                                argspec.defaults))
        required_args_count = len(argspec.args) - len(defaults)
        for arg in argspec.args[:required_args_count]:
            self.required.append(arg)
            self.properties.setdefault(arg, {'type': 'string'})
        for arg, default in defaults:
            self.properties.setdefault(arg, {
                'type': 'string',
                "default": default
            })


class operation(DocParser):
    def __init__(self, nickname='apis', **kwds):
        super(operation, self).__init__()
        self.nickname = nickname
        self.func = None
        self.func_args = []
        self.kwds = kwds

    def __call__(self, *args, **kwds):
        if self.func:
            return self.func(*args, **kwds)

        func = args[0]
        self._parse_operation(func)

        @wraps(func)
        def __wrapper__(*in_args, **in_kwds):
            return self.func(*in_args, **in_kwds)

        __wrapper__.rest_api = self
        return __wrapper__

    def _parse_operation(self, func):
        self.func = func

        self.__name__ = func.__name__
        self._parse_args(func)
        self.parse_docstring(inspect.getdoc(self.func))

    def _parse_args(self, func):
        argspec = inspect.getargspec(func)
        argspec.args.remove("self")

        defaults = []
        if argspec.defaults:
            defaults = argspec.args[-len(argspec.defaults):]

        for arg in argspec.args:
            if arg in defaults:
                required = False
            else:
                required = True
            self.params.setdefault(arg, {
                'name': arg,
                'required': required,
                'paramType': 'path',
                'dataType': 'string'
            })
        self.func_args = argspec.args


def docs(**opts):
    default_settings.update(opts)


class Application(tornado.web.Application):
    def __init__(self, handlers=None,
                 default_host="",
                 transforms=None,
                 **settings):
        super(Application, self).__init__(swagger_handlers() + handlers,
                                          default_host,
                                          transforms,
                                          **settings)