From 71e5662365892d32a5f92e2733b7ae804c833f57 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@robocracy.org>
Date: Thu, 4 Jun 2020 13:59:57 -0700
Subject: re-use 'best pdf url' for release green button

I thought this was the existing behavior, but it looks like we were just
taking the first link from the first file.

In the future may refactor this out even further.
---
 python/fatcat_web/templates/release_view.html | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'python')

diff --git a/python/fatcat_web/templates/release_view.html b/python/fatcat_web/templates/release_view.html
index 60e4624e..2aba47ba 100644
--- a/python/fatcat_web/templates/release_view.html
+++ b/python/fatcat_web/templates/release_view.html
@@ -253,8 +253,8 @@ accessible version.
 </div>
 <div class="column" style="flex: 0 0 24em;">
 
-{% if entity.state == 'active' and entity.files != [] and entity.files[0].urls != [] %}
-<a href="{{ entity.files[0].urls[0].url }}" class="ui top attached fluid huge green button"><i class="file pdf outline icon"></i>Read Full Text</a>
+{% if entity.state == 'active' and entity._es and entity._es.best_pdf_url %}
+<a href="{{ entity._es.best_pdf_url }}" class="ui top attached fluid huge green button"><i class="file pdf outline icon"></i>Read Full Text</a>
 {% elif entity.state == 'active' and entity.webcaptures != [] and entity.webcaptures[0].archive_urls != [] and entity.webcaptures[0].archive_urls[0].rel == "wayback" %}
 <a href="{{ entity.webcaptures[0].archive_urls[0].url }}{{ entity.webcaptures[0]._wayback_suffix }}" class="ui top attached fluid huge green button"><i class="file archive outline icon"></i>View Web Archive</a>
 {% elif entity.state == 'active' %}
-- 
cgit v1.2.3


From a42d5f0d00e76bf8474647fae4e1d9d61693a7d9 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@robocracy.org>
Date: Thu, 4 Jun 2020 14:01:34 -0700
Subject: ES schema: add best_url to file schema

This will increase index size (URLs are often long in our corpus, and we
have many file entities), but seems worth it.

Initially added `ia_url` as a second field, guaranteed to always be an
*.archive.org URL, but `best_url` defaults to that anyways so didn't
seem worthwhile.
---
 extra/elasticsearch/file_schema.json            |  1 +
 python/fatcat_tools/transforms/elasticsearch.py | 12 ++++++++++++
 2 files changed, 13 insertions(+)

(limited to 'python')

diff --git a/extra/elasticsearch/file_schema.json b/extra/elasticsearch/file_schema.json
index 9c8ee64c..0fa25c3a 100644
--- a/extra/elasticsearch/file_schema.json
+++ b/extra/elasticsearch/file_schema.json
@@ -44,6 +44,7 @@
             "rels":             { "type": "keyword", "normalizer": "default" },
             "in_ia":            { "type": "boolean" },
             "in_ia_petabox":    { "type": "boolean" },
+            "best_url":         { "type": "keyword", "normalizer": "default" },
 
             "release_id":       { "type": "alias", "path": "release_ids" },
             "sha1hex":          { "type": "alias", "path": "sha1" },
diff --git a/python/fatcat_tools/transforms/elasticsearch.py b/python/fatcat_tools/transforms/elasticsearch.py
index 87e054ec..1d35141b 100644
--- a/python/fatcat_tools/transforms/elasticsearch.py
+++ b/python/fatcat_tools/transforms/elasticsearch.py
@@ -508,4 +508,16 @@ def file_to_elasticsearch(entity):
     t['in_ia'] = bool('archive.org' in t['domains'])
     t['in_ia_petabox'] = bool('archive.org' in t['hosts'])
 
+    any_url = None
+    good_url = None
+    best_url = None
+    for release_url in (entity.urls or []):
+        any_url = release_url.url
+        if release_url.rel in ('webarchive', 'repository'):
+            good_url = release_url.url
+        if '//web.archive.org/' in release_url.url or '//archive.org/' in release_url.url:
+            best_url = release_url.url
+    # here is where we bake-in priority; IA-specific
+    t['best_url'] = best_url or good_url or any_url
+
     return t
-- 
cgit v1.2.3


From 2f233a3b8c00385d4b215361a0fa09f93a05f8d9 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@robocracy.org>
Date: Thu, 4 Jun 2020 14:05:56 -0700
Subject: use ES 'best_url' in file download pages

Similar to recent change for release download pages.
---
 python/fatcat_web/entity_helpers.py        | 2 ++
 python/fatcat_web/templates/file_view.html | 4 ++--
 2 files changed, 4 insertions(+), 2 deletions(-)

(limited to 'python')

diff --git a/python/fatcat_web/entity_helpers.py b/python/fatcat_web/entity_helpers.py
index d82ea0e9..e3d538e0 100644
--- a/python/fatcat_web/entity_helpers.py
+++ b/python/fatcat_web/entity_helpers.py
@@ -30,6 +30,8 @@ def enrich_creator_entity(entity):
     return entity
 
 def enrich_file_entity(entity):
+    if entity.state == "active":
+        entity._es = file_to_elasticsearch(entity)
     return entity
 
 def enrich_fileset_entity(entity):
diff --git a/python/fatcat_web/templates/file_view.html b/python/fatcat_web/templates/file_view.html
index d60ea49d..02f47a91 100644
--- a/python/fatcat_web/templates/file_view.html
+++ b/python/fatcat_web/templates/file_view.html
@@ -44,8 +44,8 @@ No known public URL, mirror, or archive for this file.
 </div>
 <div class="column" style="flex: 0 0 24em;">
 
-{% if file.urls != None and file.urls != [] %}
-<a href="{{ file.urls[0].url }}" class="ui top attached fluid huge green button"><i class="file icon"></i>Download File</a>
+{% if file._es and file._es.best_url %}
+<a href="{{ file._es.best_url }}" class="ui top attached fluid huge green button"><i class="file icon"></i>Download File</a>
 {% else %}
 <span class="ui top attached fluid huge grey button"><i class="file cross icon"></i>No Download Available</span>
 {% endif %}
-- 
cgit v1.2.3