Passed
Pull Request — master (#966)
by Konstantin
04:07 queued 01:37
created

ocrd.mets_server.ClientSideOcrdMets.add_file()   A

Complexity

Conditions 1

Size

Total Lines 12
Code Lines 12

Duplication

Lines 0
Ratio 0 %

Importance

Changes 0
Metric Value
eloc 12
dl 0
loc 12
rs 9.8
c 0
b 0
f 0
cc 1
nop 8

How to fix   Many Parameters   

Many Parameters

Methods with many parameters are not only hard to understand, but their parameters also often become inconsistent when you need more, or different data.

There are several approaches to avoid long parameter lists:

1
"""
2
# METS server functionality
3
"""
4
import re
5
from os import environ, _exit
6
from io import BytesIO
7
from typing import Any, Dict, Optional, Union, List, Tuple
8
9
from fastapi import FastAPI, Request, File, Form, UploadFile
10
from fastapi.responses import JSONResponse
11
from requests import request, Session as requests_session
12
from requests_unixsocket import Session as requests_unixsocket_session
13
from pydantic import BaseModel, Field, ValidationError
14
15
import uvicorn
16
17
from ocrd_models import OcrdMets, OcrdFile, OcrdAgent
18
from ocrd_utils import initLogging, getLogger, deprecated_alias
19
20
#
21
# XXX HACKS TODO
22
#
23
initLogging()
24
25
#
26
# Models
27
#
28
29
class OcrdFileModel(BaseModel):
30
    file_grp : str = Field()
31
    file_id : str = Field()
32
    mimetype : str = Field()
33
    page_id : Union[str, None] = Field()
34
    local_filename : str = Field()
35
36
    @staticmethod
37
    def create(file_grp : str, file_id : str, page_id : Union[str, None], local_filename : str, mimetype : str):
38
        return OcrdFileModel(file_grp=file_grp, file_id=file_id, page_id=page_id, mimetype=mimetype, local_filename=local_filename)
39
40
class OcrdAgentModel(BaseModel):
41
    name : str = Field()
42
    _type : str = Field()
43
    role : str = Field()
44
    otherrole : Optional[str] = Field()
45
    othertype : str = Field()
46
    notes : Optional[List[Tuple[Dict[str, str], Optional[str]]]] = Field()
47
48
    @staticmethod
49
    def create(name : str, _type : str, role : str, otherrole : str, othertype : str, notes : List[Tuple[Dict[str, str], Optional[str]]]):
50
        return OcrdAgentModel(name=name, _type=_type, role=role, otherrole=otherrole, othertype=othertype, notes=notes)
51
52
53
class OcrdFileListModel(BaseModel):
54
    files : List[OcrdFileModel] = Field()
55
56
    @staticmethod
57
    def create(files : List[OcrdFile]):
58
        return OcrdFileListModel(
59
            files=[OcrdFileModel.create(file_grp=f.fileGrp, file_id=f.ID, mimetype=f.mimetype, page_id=f.pageId, local_filename=f.local_filename) for f in files]
60
        )
61
62
class OcrdFileGroupListModel(BaseModel):
63
    file_groups : List[str] = Field()
64
65
    @staticmethod
66
    def create(file_groups : List[str]):
67
        return OcrdFileGroupListModel(file_groups=file_groups)
68
69
class OcrdAgentListModel(BaseModel):
70
    agents : List[OcrdAgentModel] = Field()
71
72
    @staticmethod
73
    def create(agents : List[OcrdAgent]):
74
        return OcrdAgentListModel(
75
            agents=[OcrdAgentModel(name=a.name, _type=a.type, role=a.role, otherrole=a.otherrole, othertype=a.othertype, notes=a.notes) for a in agents]
76
        )
77
78
#
79
# Client
80
#
81
82
class ClientSideOcrdFile:
83
    """
84
    Provides the same interface as :py:class:`ocrd_models.ocrd_file.OcrdFile`
85
    but without attachment to :py:class:`ocrd_models.ocrd_mets.OcrdMets` since
86
    this represents the response of the :py:class:`ocrd.mets_server.OcrdMetsServer`.
87
    """
88
89
    def __init__(self, el, mimetype=None, pageId=None, loctype='OTHER', local_filename=None, mets=None, url=None, ID=None, fileGrp=None):
90
        """
91
        Args:
92
            el (): ignored
93
        Keyword Args:
94
            mets (): ignored
95
            mimetype (string): ``@MIMETYPE`` of this ``mets:file``
96
            pageId (string): ``@ID`` of the physical ``mets:structMap`` entry corresponding to this ``mets:file``
97
            loctype (string): ``@LOCTYPE`` of this ``mets:file``
98
            local_filename (): ``@xlink:href`` of this ``mets:file`` - XXX the local file once we have proper mets:FLocat bookkeeping
99
            ID (string): ``@ID`` of this ``mets:file``
100
        """
101
        self.ID = ID
102
        self.mimetype = mimetype
103
        self.local_filename = local_filename
104
        self.loctype = loctype
105
        self.pageId = pageId
106
        self.fileGrp = fileGrp
107
108
class ClientSideOcrdAgent():
109
    """
110
    Provides the same interface as :py:class:`ocrd_models.ocrd_file.OcrdAgent`
111
    but without attachment to :py:class:`ocrd_models.ocrd_mets.OcrdMets` since
112
    this represents the response of the :py:class:`ocrd.mets_server.OcrdMetsServer`.
113
    """
114
115
    def __init__(self, el, name=None, _type=None, othertype=None, role=None, otherrole=None,
116
                 notes=None):
117
        """
118
        Args:
119
            el (): ignored
120
        Keyword Args:
121
            name (string):
122
            _type (string):
123
            othertype (string):
124
            role (string):
125
            otherrole (string):
126
            notes (dict):
127
        """
128
        self.name = name
129
        self.type = _type
130
        self.othertype = othertype
131
        self.role = role
132
        self.otherrole = otherrole
133
        self.notes = notes
134
135
class ClientSideOcrdMets():
136
    """
137
    Partial substitute for :py:class:`ocrd_models.ocrd_mets.OcrdMets` which provides for
138
    :py:meth:`ocrd_models.ocrd_mets.OcrdMets.find_files`,
139
    :py:meth:`ocrd_models.ocrd_mets.OcrdMets.find_all_files`, and
140
    :py:meth:`ocrd_models.ocrd_mets.OcrdMets.add_agent`,
141
    :py:meth:`ocrd_models.ocrd_mets.OcrdMets.agents`,
142
    :py:meth:`ocrd_models.ocrd_mets.OcrdMets.add_file` to query via HTTP a
143
    :py:class:`ocrd.mets_server.OcrdMetsServer`.
144
    """
145
146
    def __init__(self, host, port, socket):
147
        self.log = getLogger('ocrd.mets_client.%s' % ('uds' if socket else 'tcp'))
148
        if socket:
149
            self.url = f'http+unix://{socket.replace("/", "%2F")}'
150
            self.session = requests_unixsocket_session()
151
        else:
152
            self.url = f'http://{host}:{port}'
153
            self.session = requests_session()
154
155
    @deprecated_alias(ID="file_id")
156
    @deprecated_alias(pageId="page_id")
157
    @deprecated_alias(fileGrp="file_grp")
158
    def find_files(self, **kwargs):
159
        if 'pageId' in kwargs:
160
            kwargs['page_id'] = kwargs.pop('pageId')
161
        if 'ID' in kwargs:
162
            kwargs['file_id'] = kwargs.pop('ID')
163
        if 'fileGrp' in kwargs:
164
            kwargs['file_grp'] = kwargs.pop('fileGrp')
165
        r = self.session.request('GET', f'{self.url}/file', params={**kwargs})
166
        for f in r.json()['files']:
167
            yield ClientSideOcrdFile(None, ID=f['file_id'], pageId=f['page_id'], fileGrp=f['file_grp'], local_filename=f['local_filename'], mimetype=f['mimetype'])
168
169
    def find_all_files(self, *args, **kwargs):
170
        return list(self.find_files(*args, **kwargs))
171
172
    def add_agent(self, *args, **kwargs):
173
        return self.session.request('POST', f'{self.url}/agent', json=OcrdAgentModel.create(**kwargs).dict())
174
175
    @property
176
    def agents(self):
177
        return [ClientSideOcrdAgent(None, **agent_dict) for agent_dict in self.session.request('GET', f'{self.url}/agent').json()['agents']]
178
179
    @property
180
    def file_groups(self):
181
        return self.session.request('GET', f'{self.url}/file_groups').json()['file_groups']
182
183
    @deprecated_alias(pageId="page_id")
184
    @deprecated_alias(ID="file_id")
185
    def add_file(self, file_grp, content=None, file_id=None, local_filename=None, mimetype=None, page_id=None, **kwargs):
186
        return self.session.request(
187
            'POST',
188
            f'{self.url}/file',
189
            data=OcrdFileModel.create(
190
                file_id=file_id,
191
                file_grp=file_grp,
192
                page_id=page_id,
193
                mimetype=mimetype,
194
                local_filename=local_filename).dict(),
195
        )
196
197
    def save(self):
198
        self.session.request('PUT', self.url)
199
200
    def stop(self):
201
        self.session.request('DELETE', self.url)
202
203
#
204
# Server
205
#
206
207
class OcrdMetsServer():
208
209
    def __init__(self, workspace, host, port, socket):
210
        self.workspace = workspace
211
        if socket and host:
212
            raise ValueError("Expecting either socket or host/port")
213
        if not socket and not(host and port):
214
            raise ValueError("Expecting both host and port")
215
        self.host = host
216
        self.port = port
217
        self.socket = socket
218
        self.log = getLogger('ocrd.workspace_client')
219
220
    def shutdown(self):
221
        _exit(0)
222
223
    def startup(self):
224
225
        workspace = self.workspace
226
227
        app = FastAPI(
228
            title="OCR-D METS Server",
229
            description="Providing simultaneous write-access to mets.xml for OCR-D",
230
        )
231
232
        @app.exception_handler(ValidationError)
233
        async def exception_handler_validation_error(request: Request, exc: ValidationError):
234
            return JSONResponse(status_code=400, content=exc.errors())
235
236
        @app.exception_handler(FileExistsError)
237
        async def exception_handler_file_exists(request: Request, exc: FileExistsError):
238
            return JSONResponse(status_code=400, content=str(exc))
239
240
        @app.exception_handler(re.error)
241
        async def exception_handler_invalid_regex(request: Request, exc: re.error):
242
            return JSONResponse(status_code=400, content=f'invalid regex: {exc}')
243
244
        @app.get("/file", response_model=OcrdFileListModel)
245
        async def find_files(
246
            file_grp : Union[str, None] = None,
247
            file_id : Union[str, None] = None,
248
            page_id : Union[str, None] = None,
249
            mimetype : Union[str, None] = None,
250
        ):
251
            """
252
            Find files in the mets
253
            """
254
            found = workspace.mets.find_all_files(fileGrp=file_grp, ID=file_id, pageId=page_id, mimetype=mimetype)
255
            return OcrdFileListModel.create(found)
256
257
        @app.put('/')
258
        def save():
259
            return workspace.save_mets()
260
261
        @app.post('/file', response_model=OcrdFileModel)
262
        async def add_file(
263
            file_grp : str = Form(),
264
            file_id : str = Form(),
265
            page_id : Union[str, None] = Form(),
266
            mimetype : str = Form(),
267
            local_filename : str = Form(),
268
        ):
269
            """
270
            Add a file
271
            """
272
            # Validate
273
            file_resource = OcrdFileModel.create(file_grp=file_grp, file_id=file_id, page_id=page_id, mimetype=mimetype, local_filename=local_filename)
274
            # Add to workspace
275
            kwargs = file_resource.dict()
276
            kwargs['page_id'] = page_id
277
            workspace.add_file(**kwargs)
278
            return file_resource
279
280
        @app.get('/file_groups', response_model=OcrdFileGroupListModel)
281
        async def file_groups():
282
            return {'file_groups': workspace.mets.file_groups}
283
284
        @app.post('/agent', response_model=OcrdAgentModel)
285
        async def add_agent(agent : OcrdAgentModel):
286
            kwargs = agent.dict()
287
            workspace.mets.add_agent(**kwargs)
288
            return agent
289
290
        @app.get('/agent', response_model=OcrdAgentListModel)
291
        async def agents():
292
            return OcrdAgentListModel.create(workspace.mets.agents)
293
294
        @app.delete('/')
295
        async def stop():
296
            """
297
            Stop the server
298
            """
299
            getLogger('ocrd_models.ocrd_mets').info('Shutting down')
300
            workspace.save_mets()
301
            # os._exit because uvicorn catches SystemExit raised by sys.exit
302
            _exit(0)
303
304
        uvicorn.run(app, host=self.host, port=self.port, uds=self.socket)
305