Bronze

.gitignore

@@ -1,3 +1,4 @@
 *.pyc
 /.coverage
+/config.py
 /venv

.travis.yml

@@ -4,6 +4,7 @@ python:
     - "3.5"
 
 install:
+    - cp tk/default_config.py config.py
     - ./bin/build-dev
 
 script:

README.md

@@ -2,6 +2,22 @@
 
 [![Build Status](https://travis-ci.org/bartfeenstra/tk.svg?branch=master)](https://travis-ci.org/bartfeenstra/tk) [![Coverage Status](https://coveralls.io/repos/github/bartfeenstra/tk/badge.svg?branch=master)](https://coveralls.io/github/bartfeenstra/tk?branch=master)
 
+## Usage
+Substitute `http://127.0.0.1:5000` for the actual application URL, if
+you are not using `./bin/run-dev`.
+
+### Installation
+Copy `./tk/default_config.py` to `./config.py`, and override any of the
+default configuration as necessary.
+
+### Submitting a document
+`curl -X POST --header "Content-Type: application/octet-stream" --data-binary @{file_path} http://127.0.0.1:5000/submit`
+where `{file_path}` is file path of the document to process.
+
+### Retrieving a document's profile
+`curl -X GET --header "Accept: text/xml" http://127.0.0.1:5000/retrieve/{uuid}`
+where `{uuid}` is the process UUID returned by `POST /submit`.
+
 ## Development
 
 ### Building the code
@@ -14,7 +30,8 @@ Run `./bin/test`.
 Run `./bin/fix` to fix what can be fixed automatically.
 
 ### Running the application
-Run `./bin/run-dev` to start a development web server.
+Run `./bin/run-dev` to start a development web server at
+[http://127.0.0.1:5000](http://127.0.0.1:5000).
 
 ### Code style
 All code follows [PEP 8](https://www.python.org/dev/peps/pep-0008/).

bin/run-dev

@@ -2,7 +2,7 @@
 
 (
     cd `dirname "$0"`/.. &&
-    FLASK_APP=./tk/flask/entry_point.py python -m flask run
+    FLASK_APP=./tk/flask/entry_point.py TK_CONFIG_FILE=`readlink -f ./config.py` python -m flask run
 )
 
 

bin/test

@@ -3,6 +3,6 @@
 (
     cd `dirname "$0"`/..
     flake8 --ignore=E501 ./tk
-    coverage run -m nose2 &&
+    TK_CONFIG_FILE=`readlink -f ./config.py` coverage run -m nose2 &&
     coverage report -m
 )

requirements-dev-frozen.txt

@@ -5,17 +5,19 @@ click==6.7
 coverage==4.4.2
 coveralls==1.2.0
 docopt==0.6.2
-flake8==3.3.0
+flake8==3.5.0
 Flask==0.12.2
 idna==2.6
 itsdangerous==0.24
 Jinja2==2.10
 MarkupSafe==1.0
 mccabe==0.6.1
-nose2==0.6.5
+nose2==0.7.2
 pycodestyle==2.3.1
-pyflakes==1.5.0
+pyflakes==1.6.0
 requests==2.18.4
+requests-futures==0.9.7
+requests-mock==1.3.0
 six==1.11.0
 urllib3==1.22
 Werkzeug==0.12.2

requirements-dev.txt

@@ -1,6 +1,7 @@
 -r ./requirements.txt
 autopep8 ~= 1.3
-coverage ~= 4.4.1
+coverage ~= 4.4.2
 coveralls ~= 1.2.0
-flake8 ~= 3.3.0
-nose2 ~= 0.6.5
+flake8 ~= 3.5.0
+nose2 ~= 0.7.2
+requests-mock ~= 1.3.0

requirements-frozen.txt

@@ -1,6 +1,12 @@
+certifi==2017.11.5
+chardet==3.0.4
 click==6.7
 Flask==0.12.2
+idna==2.6
 itsdangerous==0.24
 Jinja2==2.10
 MarkupSafe==1.0
+requests==2.18.4
+requests-futures==0.9.7
+urllib3==1.22
 Werkzeug==0.12.2

requirements.txt

@@ -1 +1,3 @@
-flask ~= 0.12.2
+flask ~= 0.12.2
+requests ~= 2.18.4
+requests-futures ~= 0.9.7

tk/default_config.py

@@ -0,0 +1,4 @@
+SOURCEBOX_URL = 'https://staging.textkernel.nl/sourcebox/extract.do'
+SOURCEBOX_ACCOUNT_NAME = None
+SOURCEBOX_USER_NAME = None
+SOURCEBOX_PASSWORD = None

tk/flask/app.py

@@ -1,5 +1,76 @@
-from flask import Flask
+from flask import Flask, request, Response
+from requests_futures.sessions import FuturesSession
+from werkzeug.exceptions import NotAcceptable, UnsupportedMediaType, NotFound, \
+    BadRequest
+
+from tk.process import Process
+
+
+def request_content_type(content_type):
+    """
+    Check we can accept the request body.
+    :param content_type:
+    :return:
+    """
+
+    def decorator(route_method):
+        def checker(*route_method_args, **route_method_kwargs):
+            if request.mimetype != content_type:
+                raise UnsupportedMediaType()
+            return route_method(*route_method_args, **route_method_kwargs)
+
+        return checker
+
+    return decorator
+
+
+def response_content_type(content_type):
+    """
+    Check we can deliver the right content type.
+    :return:
+    """
+
+    def decorator(route_method):
+        def checker(*route_method_args, **route_method_kwargs):
+            negotiated_content_type = request.accept_mimetypes.best_match(
+                [content_type])
+            if negotiated_content_type is None:
+                raise NotAcceptable()
+            return route_method(*route_method_args, **route_method_kwargs)
+
+        return checker
+
+    return decorator
 
 
 class App(Flask):
-    pass
+    def __init__(self, *args, **kwargs):
+        super().__init__('tk', static_folder=None, *args, **kwargs)
+        self.config.from_object('tk.default_config')
+        self.config.from_envvar('TK_CONFIG_FILE')
+        self._register_routes()
+        self._session = FuturesSession()
+        self._process = Process(self._session, self.config['SOURCEBOX_URL'],
+                                self.config['SOURCEBOX_ACCOUNT_NAME'],
+                                self.config['SOURCEBOX_USER_NAME'],
+                                self.config['SOURCEBOX_PASSWORD'])
+
+    def _register_routes(self):
+        @self.route('/submit', methods=['POST'], endpoint='submit')
+        @request_content_type('application/octet-stream')
+        @response_content_type('text/plain')
+        def submit():
+            document = request.get_data()
+            if not document:
+                raise BadRequest()
+            process_id = self._process.submit(document)
+            return Response(process_id, 200, mimetype='text/plain')
+
+        @self.route('/retrieve/<process_id>', endpoint='retrieve')
+        @request_content_type('')
+        @response_content_type('text/xml')
+        def retrieve(process_id):
+            result = self._process.retrieve(process_id)
+            if result is None:
+                raise NotFound()
+            return Response(result, 200, mimetype='text/plain')

tk/flask/entry_point.py

@@ -1,3 +1,3 @@
 from tk.flask.app import App
 
-app = App('tk')
+app = App()

tk/process.py

@@ -0,0 +1,52 @@
+import uuid
+from queue import Queue
+from threading import Thread
+
+
+class Process:
+    def __init__(self, session, sourcebox_url, sourcebox_account_name, sourcebox_user_name, sourcebox_password):
+        self._processes = {}
+        self._process_queue = Queue()
+        self._session = session
+        self._sourcebox_url = sourcebox_url
+        self._sourcebox_account_name = sourcebox_account_name
+        self._sourcebox_user_name = sourcebox_user_name
+        self._sourcebox_password = sourcebox_password
+
+        worker = Thread(target=self._process_queue_worker,
+                        args=(self._process_queue,))
+        worker.setDaemon(True)
+        worker.start()
+
+    def _process_queue_worker(self, queue):
+        while True:
+            process_id, profile = queue.get()
+            self._processes[process_id] = profile
+            queue.task_done()
+
+    def submit(self, document):
+        process_id = str(uuid.uuid4())
+        self._processes[process_id] = 'PROGRESS'
+        self._session.post(self._sourcebox_url, data={
+            'account': self._sourcebox_account_name,
+            'username': self._sourcebox_user_name,
+            'password': self._sourcebox_password,
+        }, files={
+            'uploaded_file': document,
+        }, params={
+            'useHttpErrorCodes': 'true',
+            'useJsonErrorMsg': 'true',
+        }, background_callback=self._handle_submit_response(process_id))
+        return process_id
+
+    def _handle_submit_response(self, process_id):
+        queue = self._process_queue
+
+        def _handler(session, response):
+            queue.put((process_id, response.text))
+        return _handler
+
+    def retrieve(self, process_id):
+        if process_id not in self._processes:
+            return None
+        return self._processes[process_id]

tk/tests/__init__.py

@@ -0,0 +1,74 @@
+from unittest import TestCase
+
+import requests
+import requests_mock
+
+from tk.flask.app import App
+
+
+def data_provider(data_provider):
+    """
+    Provides a test method with test data.
+
+    Applying this decorator to a test method causes that method to be run as
+    many times as the data provider callable returns dictionary items.
+    Failed assertions will include information about which data set failed.
+
+    :param data_provider: A callable that generates the test data as a
+      dictionary of tuples containing the test method arguments, keyed by data
+      set name.
+    :return:
+    """
+    def decorator(test_method):
+        """
+        The actual decorator.
+        :param test_method: The test method to decorate.
+        :return:
+        """
+
+        def multiplier(self, *test_method_args, **test_method_kwargs):
+            """
+            The replacement (decorated) test method.
+            :param self:
+            :param test_method_args: The arguments to the decorated test
+             method.
+            :param test_method_kwargs: The keyword arguments to the decorated
+             test method.
+            :return:
+            """
+            for fixture_name, test_method_fixture_args in data_provider().items():
+                try:
+                    test_method(self, *test_method_args,
+                                *test_method_fixture_args,
+                                **test_method_kwargs)
+                except AssertionError:
+                    raise AssertionError(
+                        'Assertion failed with data set "%s"' % str(fixture_name))
+                except Exception:
+                    raise AssertionError(
+                        'Unexpected error with data set "%s"' % str(
+                            fixture_name))
+
+        return multiplier
+
+    return decorator
+
+
+class IntegrationTestCase(TestCase):
+    """
+    Provides scaffolding for light-weight integration tests.
+    """
+
+    def setUp(self):
+        self._flask_app = App()
+        self._flask_app.config.update(SERVER_NAME='example.com')
+        self._flask_app_context = self._flask_app.app_context()
+        self._flask_app_context.push()
+        self._flask_app_client = self._flask_app.test_client()
+
+        session = requests.Session()
+        adapter = requests_mock.Adapter()
+        session.mount('mock', adapter)
+
+    def tearDown(self):
+        self._flask_app_context.pop()