Development

api.config.example

@@ -1,6 +1,9 @@
 # Rename this file to "api.config" to make it active
 
 [apis]
+api_version = 3
+default_apis_base_url = {{ apis_common_base_URL }}
+production_apis_base_url = {{ apis_production_base_URL }}
 
 # if true, blocks all writing behavior
 read_only = READ_ONLY_MODE
@@ -37,11 +40,12 @@ following_repo_remote   = FOLLOWING_REPO_REMOTE
 git_ssh     = GIT_SSH
 pkey        = PKEY
 
-oti_base_url = OTI_BASE_URL
-otindex_base_url = OTINDEX_BASE_URL 
-collections_api_base_url = COLLECTIONS_API_BASE_URL
-amendments_api_base_url = AMENDMENTS_API_BASE_URL
-favorites_api_base_url = FAVORITES_API_BASE_URL
+# deprecate these overly-specific base URLs? or copy same value to all?
+##oti_base_url = OTI_BASE_URL
+##otindex_base_url = OTINDEX_BASE_URL 
+##collections_api_base_url = COLLECTIONS_API_BASE_URL
+##amendments_api_base_url = AMENDMENTS_API_BASE_URL
+##favorites_api_base_url = FAVORITES_API_BASE_URL
 
 opentree_docstore_url = OPENTREE_DOCSTORE_URL
 # Push scripts will substitute the OTI_BASE_URL server-config variables, shared

phylesystem_api/phylesystem_api/api_utils.py

@@ -43,7 +43,7 @@
 READ_ONLY_MODE = True
 
 def get_private_dir(request):
-    _LOG.debug("WHY PROVATE DIR")
+    _LOG.debug("WHY PRIVATE DIR")
     return "~/private/"
 
 def atomic_write_json_if_not_found(obj, dest, request):
@@ -669,7 +669,7 @@ def find_in_request(request, property_name, default_value=None, return_all_value
 # N.B. HTML comments are stripped by default. Non-allowed tags will appear
 # "naked" in output, so we can identify any bad actors.
 allowed_curation_comment_tags = ['p', 'br', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'hr', 'pre', 'code']  # any others?
-ot_markdown_tags = list(set( bleach.sanitizer.ALLOWED_TAGS + allowed_curation_comment_tags))
+ot_markdown_tags = list(set(bleach.sanitizer.ALLOWED_TAGS)) + list(set(allowed_curation_comment_tags))
 # allow hyperlinks with target="_blank"
 ot_markdown_attributes = {}
 ot_markdown_attributes.update(bleach.sanitizer.ALLOWED_ATTRIBUTES)

phylesystem_api/phylesystem_api/routes.py

@@ -16,7 +16,6 @@ def includeme(config):
     config.add_route('merge_docstore_changes', '/{api_version}/merge_docstore_changes/{doc_id}/{starting_commit_SHA}')
     config.add_route('push_docstore_changes', '/{api_version}/push_docstore_changes/{doc_type}/{doc_id}')
     config.add_route('push_docstore_changes_bare', '/{api_version}/push_docstore_changes',  request_method='PUT')
-
     #
     # STUDY/TREE ROUTES
     #
@@ -29,12 +28,15 @@ def includeme(config):
     config.add_route('create_study', '/{api_version}/study/')
     config.add_route('study_CORS_preflight', '/{api_version}/study/{study_id}', request_method='OPTIONS')
     config.add_route('fetch_study', '/{api_version}/study/{study_id}', request_method='GET')
+    config.add_route('fetch_study_label', '/{api_version}/study/{study_id}/', request_method='GET')
     config.add_route('update_study', '/{api_version}/study/{study_id}', request_method='PUT')
     config.add_route('delete_study', '/{api_version}/study/{study_id}', request_method='DELETE')
     config.add_route('get_study_file_list', '/{api_version}/study/{study_id}/file')
     config.add_route('get_study_single_file', '/{api_version}/study/{study_id}/file/{file_id}')
     config.add_route('get_study_external_url', '/{api_version}/study/external_url/{study_id}')
     config.add_route('get_study_tree', '/{api_version}/study/{study_id}/tree/{tree_id_with_extension}')
+    config.add_route('get_study_tree_label', '/{api_version}/study/{study_id}/tree/{tree_id_with_extension}/')
+
     #
     # TREE COLLECTION ROUTES
     #

phylesystem_api/phylesystem_api/views/default.py

@@ -67,6 +67,15 @@ def base_API_view(request):
         "source_url": "https://github.com/OpenTreeOfLife/phylesystem-api"
     }
 
+
+
+# Create a unique cache key with the URL and any vars (GET *and* POST) to its "query string"
+# ALSO include the request method (HTTP verb) to respond to OPTIONS requests
+def create_unique_cache_key(target_url, request):
+    unique_key = "cached:{}:{}:{}".format(request.method, target_url, request.body.decode('utf-8'))
+    #_LOG.warn(">> unique cache key: {}".format(unique_key))
+    return unique_key
+
 @view_config(route_name='pull_through_cache')
 def pull_through_cache(request):
     """
@@ -77,23 +86,46 @@ def pull_through_cache(request):
     """
 #    _LOG = api_utils.get_logger(request, 'ot_api')
     api_utils.raise_on_CORS_preflight(request)
+
+    # gather any request elements used to build a unique cache key
     target_url = request.matchdict.get('target_url')
     _LOG.warn(">> target_url: {}".format(target_url))
 
-    @cache_region('short_term', 'pull-through')
+    # Some headers should not be used when adding to our RAM cache
+    hop_by_hop_headers = ['Keep-Alive',
+                          'Transfer-Encoding',
+                          'TE',
+                          'Connection',
+                          'Trailer',
+                          'Upgrade',
+                          'Proxy-Authorization',
+                          'Proxy-Authenticate',
+                          ]
+
+    @cache_region('short_term', create_unique_cache_key(target_url, request))
     def fetch_and_cache(url):
         # let's restrict this to URLs on this api server, to avoid shenanigans
         #import pdb; pdb.set_trace()
         root_relative_url = "/{}".format(url)
         _LOG.warn(">> root_relative_url: {}".format(root_relative_url))
-        fetch_url = request.relative_url(root_relative_url)
+        conf = api_utils.get_conf_object(request)
+        base_url = conf.get("apis", "default_apis_base_url")
+        fetch_url =  base_url + root_relative_url
         _LOG.warn("NOT CACHED, FETCHING THIS URL: {}".format(fetch_url))
         _LOG.warn("  request.method = {}".format(request.method))
+
+        # modify or discard "hop-by-hop" headers
+        for bad_header in hop_by_hop_headers:
+            request.headers.pop(bad_header, None)
+        #_LOG.warn("  MODIFIED request.headers:")
+        #_LOG.warn( dict(request.headers) )
+
         try:
             if request.method == 'POST':
                 # assume a typical API request with JSON payload
                 # (pass this along unchanged)
                 _LOG.warn("  treating as POST")
+                _LOG.warn("  headers: {}".format(request.headers))
                 fetched = requests.post(url=fetch_url,
                                         data=request.body,
                                         headers=request.headers)
@@ -110,6 +142,13 @@ def fetch_and_cache(url):
             _LOG.warn("... and now we're back with fetched, which is a {}".format( type(fetched) ))
             fetched.raise_for_status()
             fetched.encoding = 'utf-8' # Optional: requests infers this internally
+
+            # modify or discard "hop-by-hop" headers
+            for bad_header in hop_by_hop_headers:
+                fetched.headers.pop(bad_header, None)
+            #_LOG.warn("  MODIFIED fetched.headers:")
+            #_LOG.warn( dict(fetched.headers) )
+
             try:
                 test_for_json = fetched.json()  # missing JSON payload will raise an error
                 return Response(

phylesystem_api/phylesystem_api/views/study.py

@@ -162,22 +162,23 @@ def __finish_write_verb(phylesystem,
 
 
 
-
-@view_config(route_name='fetch_study', renderer=None)
+@view_config(route_name='fetch_study', renderer=None, request_method='GET')
+@view_config(route_name='fetch_study_label', renderer=None, request_method='GET')
 def fetch_study(request):
     repo_parent, repo_remote, git_ssh, pkey, git_hub_remote, max_filesize, max_num_trees, read_only_mode = api_utils.read_phylesystem_config(request)
     #_LOG = api_utils.get_logger(request, 'ot_api.default.v1')
     _LOG.debug("Fetching study")
     api_version = request.matchdict['api_version']
     study_id = request.matchdict['study_id']
+    _LOG.debug('study_id = {}'.format(study_id))
     content_id = None
     version_history = None
     comment_html = None
-    final_path_part = request.path.split('/')[-1] ##TODO What if there are other parts...
     # does this look like a filename? if so, grab its extension
     request_extension = None
-    fpps = final_path_part.split('.')
+    fpps = study_id.split('.')
     if len(fpps) > 1:
+        _LOG.debug('len(fpps) > 1')
         request_extension = fpps[-1]
         study_id = '.'.join(fpps[:-1])
         _LOG.debug("Request extension is {}".format)
@@ -208,7 +209,7 @@ def fetch_study(request):
         except:
             comment_html = ''
     except:
-        # _LOG.exception('GET failed')
+        _LOG.exception('GET failed')
         e = sys.exc_info()[0]
         raise HTTPBadRequest(e)
 
@@ -692,7 +693,8 @@ def get_study_external_url(request):
     except:
         raise HTTPNotFound(body='{"error": 1, "description": "study not found"}')
 
-@view_config(route_name='get_study_tree', renderer='json')
+@view_config(route_name='get_study_tree', renderer=None)
+@view_config(route_name='get_study_tree_label', renderer=None)
 def get_study_tree(request):
     api_utils.raise_on_CORS_preflight(request)
 
@@ -749,4 +751,8 @@ def get_study_tree(request):
     if result_data is None:
         raise HTTPNotFound(body='subresource "tree/{t}" not found in study "{s}"'.format(t=tree_id,
                                                                                  s=study_id))
-    return result_data
+    if out_schema.is_json():
+        return render_to_response('json', result_data, request)
+    else:
+        # _LOG.debug(result_data)
+        return render_to_response('string', result_data, request)