S3: Content type detection

+ Rework content type detection as current way leads
  into inconsistent detection and corrupted CloudFront handling
+ MimeType is base on wild guess of file extension and does not work
  for our use cases.

Signed-off-by: pmikus <pmikus@cisco.com>
Change-Id: If8a3e052a5512dca1767bccf10107c9867ba7366

1 files changed, 44 insertions, 29 deletions

diff --git a/jjb/scripts/publish_library_py.sh b/jjb/scripts/publish_library_py.sh
index 400f77e42..1cbeb23c0 100644
--- a/jjb/scripts/publish_library_py.sh
+++ b/jjb/scripts/publish_library_py.sh
@@ -35,7 +35,6 @@ import shutil
 import subprocess
 import sys
 import tempfile
-from mimetypes import MimeTypes
 
 import boto3
 from botocore.exceptions import ClientError
@@ -51,6 +50,29 @@ logging.basicConfig(
 logging.getLogger(u"botocore").setLevel(logging.INFO)
 
 
+FILE_TYPE = {
+    u"xml": u"application/xml",
+    u"html": u"text/html",
+    u"txt": u"text/plain",
+    u"log": u"text/plain",
+    u"css": u"text/css",
+    u"md": u"text/markdown",
+    u"rst": u"text/x-rst",
+    u"csv": u"text/csv",
+    u"svg": u"image/svg+xml",
+    u"jpg": u"image/jpeg",
+    u"png": u"image/png",
+    u"gif": u"image/gif",
+    u"js": u"application/javascript",
+    u"pdf": u"application/pdf",
+    u"json": u"application/json",
+    u"otf": u"font/otf",
+    u"ttf": u"font/ttf",
+    u"woff": u"font/woff",
+    u"woff2": u"font/woff2"
+}
+
+
 def compress_text(src_dpath):
     """Compress all text files in directory.
 
@@ -129,35 +151,28 @@ def upload(s3_resource, s3_bucket, src_fpath, s3_path):
     :type src_fpath: str
     :type s3_path: str
     """
-    mime = MimeTypes().guess_type(src_fpath)[0]
-    encoding = MimeTypes().guess_type(src_fpath)[1]
-
-    extra_args = {
-        u"ContentType": u"text/plain"
-    }
-    text_html_extra_args = {
-        u"ContentType": u"text/html"
-    }
-    text_plain_extra_args = {
-        u"ContentType": u"text/plain"
-    }
-    app_xml_extra_args = {
-        u"ContentType": u"application/xml"
-    }
-
-    if mime is None and encoding is None:
-        extra_args = extra_args
-    elif mime is None or mime in u"text/plain":
-        extra_args = text_plain_extra_args
-    elif mime in u"text/html":
-        extra_args = text_html_extra_args
-    elif mime in u"application/xml":
-        extra_args = app_xml_extra_args
-    else:
-        extra_args = extra_args
+    def is_gzip_file(filepath):
+        with open(filepath, u"rb") as test_f:
+            return test_f.read(2) == b"\x1f\x8b"
+
+    if os.path.isdir(src_fpath):
+        return
+    if os.path.isfile(src_fpath):
+        file_name, file_extension = os.path.splitext(src_fpath)
+        content_encoding = u""
+        content_type = u"application/octet-stream"
+        if is_gzip_file(src_fpath):
+            file_name, file_extension = os.path.splitext(file_name)
+            content_encoding = "gzip"
+        content_type = FILE_TYPE.get(
+            file_extension.strip("."),
+            u"application/octet-stream"
+        )
 
-    if encoding:
-        extra_args[u"ContentEncoding"] = encoding
+        extra_args = dict()
+        extra_args[u"ContentType"] = content_type
+        if content_encoding:
+            extra_args[u"ContentEncoding"] = content_encoding
 
     try:
         s3_resource.Bucket(s3_bucket).upload_file(