From 09df0e72a1b3f72b61f53f9871b505a608efe81c Mon Sep 17 00:00:00 2001
From: Elias Gabriel <elias.gabriel@indico.io>
Date: Tue, 27 Jul 2021 10:57:54 -0700
Subject: [PATCH] [CAT-103] FIX: Support gzip deserialization (#116)

* ADD: gzip as a valid MIME type

* REMOVE: gzip decompression check by extension
---
 indico/http/client.py                         |  21 ++++++++----------
 indico/http/serialization.py                  |  15 +++++++------
 tests/unit/http/mocked_data/application_gzip_ | Bin 0 -> 1183 bytes
 tests/unit/http/test_unit_serialization.py    |  15 +++++++++----
 4 files changed, 28 insertions(+), 23 deletions(-)
 create mode 100644 tests/unit/http/mocked_data/application_gzip_

diff --git a/indico/http/client.py b/indico/http/client.py
index 08c6dbb3..900b7ef0 100644
--- a/indico/http/client.py
+++ b/indico/http/client.py
@@ -1,16 +1,17 @@
-import logging
-import requests
 import http.cookiejar
-from pathlib import Path
-from contextlib import contextmanager
+import logging
 from collections import defaultdict
+from contextlib import contextmanager
+from copy import deepcopy
+from pathlib import Path
 from typing import Union
+
+import requests
+from indico.client.request import HTTPRequest
 from indico.config import IndicoConfig
+from indico.errors import IndicoAuthenticationFailed, IndicoRequestError
 from indico.http.serialization import deserialize
-from indico.errors import IndicoRequestError, IndicoAuthenticationFailed
-from indico.client.request import HTTPRequest
 from requests import Response
-from copy import deepcopy
 
 logger = logging.getLogger(__file__)
 
@@ -128,14 +129,10 @@ def _make_request(
         # code, api_response =
         url_parts = path.split(".")
         json = False
-        gzip = False
         if len(url_parts) > 1 and (url_parts[-1] == "json" or url_parts[-2] == "json"):
             json = True
 
-        if url_parts[-1] == "gz":
-            gzip = True
-
-        content = deserialize(response, force_json=json, gzip=gzip)
+        content = deserialize(response, force_json=json)
 
         # If auth expired refresh
         if response.status_code == 401 and not _refreshed:
diff --git a/indico/http/serialization.py b/indico/http/serialization.py
index 23f8046e..a89f90e7 100644
--- a/indico/http/serialization.py
+++ b/indico/http/serialization.py
@@ -1,15 +1,15 @@
 """
 Handles deserialization / decoding of responses
 """
-from collections import defaultdict
 import cgi
-import json
-import io
 import gzip
+import io
+import json
 import logging
-import msgpack
 import traceback
+from collections import defaultdict
 
+import msgpack
 from indico.errors import IndicoDecodingError
 
 logger = logging.getLogger(__name__)
@@ -25,11 +25,10 @@ def raw_bytes(content, *args, **kwargs):
     return content
 
 
-def deserialize(response, gzip=False, force_json=False):
+def deserialize(response, force_json=False):
     content_type, params = cgi.parse_header(response.headers.get("Content-Type"))
 
-    content = None
-    if gzip:
+    if content_type in ["application/x-gzip", "application/gzip"]:
         content = decompress(response)
     else:
         content = response.content
@@ -85,5 +84,7 @@ def zip_serialization(content, charset=None):
         "image/png": image_serialization,
         "image/jpg": image_serialization,
         "application/zip": zip_serialization,
+        "application/x-gzip": raw_bytes,
+        "application/gzip": raw_bytes,
     },
 )
diff --git a/tests/unit/http/mocked_data/application_gzip_ b/tests/unit/http/mocked_data/application_gzip_
new file mode 100644
index 0000000000000000000000000000000000000000..3e403e4d13ff501f5573b30dff476babd882f5e8
GIT binary patch
literal 1183
zcmV;Q1Yr9giwFP!000001MOGqZre5#&Tj#FhXVrv5+I={>gu2<kSlg+njkSY3~5j>
zGHol6C6A<%v_+qBKlTROi|kN}?bLQtXG_u!)d#~O56N?=Lw+AAJ{o0d!gX5YZ!Z=8
ztvo&`A`E@=bfbXYeO6%1UvYivhaQ%xw?bXd_vs4ICz#6|N~O63SP@yu)jj>b)c!xk
zhgMQ_MQaOu>TUp5axHR=^$#DhDx@sr+a%*kp{aE$<W-u8{hLIL%#LYIy6AT!<D~Vf
z=%@LK_`OV}=$huCP>T1bZICLEx9y+{;7Ni#Ko5+)D77H1k6$|{-xK&O|4ZH5j2}0~
z!};%fXlnTX3~2cOed8|wN8R329RFY9e-O=||14k)|DOY^$9o5xF7a$@^{>DG_|vv*
z%LNL>#ieb%cwt*k2bm4dF4sIO`dB|Ib(ZEr0mE)5XgEfCEAZ;t%a^t#@@_TUw%7ts
z<MC+NEy`R2ogL*Z7*z$C3}IOMIqfPq1K-4ju=+oTSy%pNTvgMSD0Jg_J10UFr9>@)
zs_Oyj;BCo;cICbH?HtP@ITae>=?goDqMP!a;s(wrT0@r+v#%j)*M;+H;+3!}2o@zO
zn9dWNS0eB00k|~s!E{qv3O+R1{K>X<I<{qkP>}t4cF36sIz4m}5f5nv5o3hK7{AlC
ztu@$&({wn>1mH@;L5hWS=wPV*PVKGs;nK=S*d{E%p<x`8kj7&l9V5^9<DiB-%b+WU
zMUL~TbOQ(q2lR~R&82*YK8dJ5_I-&AgkB+`R08mT&;tR#q*9kWgF}%Fc%CMBK$N<W
zDbEyv!$Jzs;zkqL&bu%))n&p4j6?w&hn;b-Om&$cEGDr($=ie6f3CVrST`hbJG}SZ
zb})Z=e*@kdEE9a0xYQ?Yf3lSS`Bj(BbAZ%px7l)eUdA65X#DK!79eL`@BVe0sqSBi
zys8s0nv&)vUZjwOBa}7NWJ1U^E(*+8Jf>O`etVi2Ez-njfhLAg_!yEnD~Rl0NFpXW
zndHAGi($P2E`i@L6OXDXQAolde0ELwRHkbOM}rL*6xVQ3i3#Afzj~PCLFuyBLzyWx
z<b?*V#+|Tq&2=&`<%jF^h^M&iy7JwVYf9)-64S{hsr8aTb>3{?(U+xK<7T<zw1Y<#
zZdh!@VAVNh-c@AT-FOR07>_sinPYaL(?NgdcRREfu+$I`E2rg!uIk3>qTw|knpRWX
z;Yw6Ib$$@pmC$&#N9%r$A9*JI&aY{{ovU<gnB|;HKMA{w)5c>L)=v19fH!Lctw&pq
zbA5eHa&aY#oFfN@j>OhC>+7`;bjw8Gh}U&H><N|jbG1J9hWCX%t_7^^Z87{RpdPM5
z#9|h%&(7sWik>01i|<7z!0i!u0raLe7gJTW<WpPQo3puoG-qRs&Dp#*owd2ad_Omw
z@5g*L7nk`#IA;s+W;e0f7T%*!h2(A0Pt*6JdTHmRKpwfbV*F#?E1+7RTctEA$fd5+
x#3I+nEBxyI(SdD!?Qe<(4H`6P(4aws1`QfCXwaZRg9hIY{sEZf-u?h6000qWOr8J$

literal 0
HcmV?d00001

diff --git a/tests/unit/http/test_unit_serialization.py b/tests/unit/http/test_unit_serialization.py
index ed64a42b..fca9bc2c 100644
--- a/tests/unit/http/test_unit_serialization.py
+++ b/tests/unit/http/test_unit_serialization.py
@@ -1,11 +1,10 @@
-from unittest.mock import MagicMock
-from pathlib import Path
 import logging
+from pathlib import Path
+from unittest.mock import MagicMock
 
 import pytest
-
-from indico.http.serialization import deserialize
 from indico.errors import IndicoDecodingError
+from indico.http.serialization import deserialize
 
 
 @pytest.fixture(scope="function")
@@ -26,6 +25,7 @@ def _mock_loader(mime, charset):
         with open(path, "rb") as f:
             content = f.read()
         response_mock.content = content
+        response_mock.raw.data = content
         return response_mock
 
     return _mock_loader
@@ -66,6 +66,13 @@ def test_deserialize_pdf(mock_loader):
     assert isinstance(content, bytes)
 
 
+def test_deserialize_gzip(mock_loader):
+    response = mock_loader("application/gzip", "")
+    content = deserialize(response)
+
+    assert isinstance(content, bytes)
+
+
 def test_deserialize_unknown(mock_loader):
     response = mock_loader("unknown", "")
     try: