added endpoint to clean file and return it directly for automated clients

3 files changed, 137 insertions, 3 deletions

diff --git a/matweb/oas/remove_metadata.yml b/matweb/oas/remove_metadata.yml
new file mode 100644
index 0000000..ad5a7da
--- /dev/null
+++ b/matweb/oas/remove_metadata.yml
@@ -0,0 +1,41 @@
+---
+tags:
+  - "Metadata removal cleaning in one request (automated clients)"
+summary: 'Upload a single file which will be cleaned from metadata and returned directly'
+requestBody:
+  description: "The file that will be cleaned from metadata and the cleaned file is returned directly"
+  required: true
+  content:
+    multipart/form-data:
+      schema:
+        type: object
+        properties:
+          file:
+            type: string
+            format: binary
+responses:
+  '200':
+    description: "The cleaned file"
+    content:
+      "*/*":
+        schema:
+          type: string
+          format: binary
+  400:
+    description: "Invalid input"
+    content:
+      application/json:
+        schema:
+          $ref: '#/components/schemas/ErrorResponse'
+  415:
+    description: "Unsupported file type"
+    content:
+      application/json:
+        schema:
+          $ref: '#/components/schemas/ErrorResponse'
+  500:
+    description: "Unable to clean the file"
+    content:
+      application/json:
+        schema:
+          $ref: '#/components/schemas/ErrorResponse'
\ No newline at end of file
diff --git a/matweb/rest_api.py b/matweb/rest_api.py
index 2909cf6..50b7c37 100644
--- a/matweb/rest_api.py
+++ b/matweb/rest_api.py
@@ -84,6 +84,38 @@ class APIDownload(Resource):
         return send_from_directory(current_app.config['UPLOAD_FOLDER'], filepath, as_attachment=True)
 
 
+class APIClean(Resource):
+    @swag_from('./oas/remove_metadata.yml')
+    def post(self):
+        if 'file' not in request.files:
+            abort(400, message='No file part')
+
+        uploaded_file = request.files['file']
+        if not uploaded_file.filename:
+            abort(400, message='No selected `file`')
+        try:
+            filename, filepath = utils.save_file(uploaded_file, current_app.config['UPLOAD_FOLDER'])
+        except ValueError:
+            abort(400, message='Invalid Filename')
+
+        parser, mime = utils.get_file_parser(filepath)
+
+        if parser is None:
+            abort(415, message='The type %s is not supported' % mime)
+
+        if parser.remove_all() is not True:
+            abort(500, message='Unable to clean %s' % mime)
+
+        _, _, _, output_filename = utils.cleanup(parser, filepath, current_app.config['UPLOAD_FOLDER'])
+
+        @after_this_request
+        def remove_file(response):
+            os.remove(os.path.join(current_app.config['UPLOAD_FOLDER'], output_filename))
+            return response
+
+        return send_from_directory(current_app.config['UPLOAD_FOLDER'], output_filename, as_attachment=True)
+
+
 class APIBulkDownloadCreator(Resource):
     schema = {
         'download_list': {
@@ -169,6 +201,10 @@ api.add_resource(
     '/download/<string:key>/<string:secret>/<string:filename>'
 )
 api.add_resource(
+    APIClean,
+    '/remove_metadata'
+)
+api.add_resource(
     APIBulkDownloadCreator,
     '/download/bulk'
 )
diff --git a/test/test_api.py b/test/test_api.py
index 6961197..6d48a35 100644
--- a/test/test_api.py
+++ b/test/test_api.py
@@ -1,13 +1,15 @@
-import unittest
-import tempfile
 import json
 import os
 import shutil
+import tempfile
+import unittest
 import zipfile
-from six import BytesIO
+import io
 
 from unittest.mock import patch
+
 from openapi_spec_validator import validate_spec
+from six import BytesIO
 
 import main
 
@@ -433,5 +435,60 @@ class Mat2APITestCase(unittest.TestCase):
         spec = self.app.get('apispec_1.json').get_json()
         validate_spec(spec)
 
+    def test_remove_metadata(self):
+        r = self.app.post(
+            '/api/remove_metadata',
+            data=dict(
+                file=(io.BytesIO(b""), 'test.txt'),
+            ),
+            follow_redirects=False
+        )
+        self.assertEqual(r.status_code, 200)
+        self.assertEqual(r.headers['Content-Disposition'], 'attachment; filename=test.cleaned.txt')
+        self.assertEqual(r.headers['Content-Type'], 'text/plain; charset=utf-8')
+        self.assertEqual(r.data, b'')
+
+    def test_remove_metdata_validation(self):
+        r = self.app.post(
+            '/api/remove_metadata',
+            data=dict(
+                fileNotExisting=(io.BytesIO(b""), 'test.random'),
+            ),
+            follow_redirects=False
+        )
+        self.assertEqual(r.get_json()['message'], 'No file part')
+        self.assertEqual(r.status_code, 400)
+
+        r = self.app.post(
+            '/api/remove_metadata',
+            data=dict(
+                file=(io.BytesIO(b""), ''),
+            ),
+            follow_redirects=False
+        )
+        self.assertEqual(r.get_json()['message'], 'No selected `file`')
+        self.assertEqual(r.status_code, 400)
+
+        r = self.app.post(
+            '/api/remove_metadata',
+            data=dict(
+                file=(io.BytesIO(b""), '../../'),
+            ),
+            follow_redirects=False
+        )
+        self.assertEqual(r.get_json()['message'], 'Invalid Filename')
+        self.assertEqual(r.status_code, 400)
+
+        r = self.app.post(
+            '/api/remove_metadata',
+            data=dict(
+                file=(io.BytesIO(b""), 'test.random'),
+            ),
+            follow_redirects=False
+        )
+        self.assertEqual(r.get_json()['message'], 'The type None is not supported')
+        self.assertEqual(r.status_code, 415)
+
+
 if __name__ == '__main__':
     unittest.main()