Downrank canon tag matches

Author: khemarato

_content/articles/lao-buddhist-women_tsomo-karma-lekshe.md

@@ -11,7 +11,6 @@ course: theravada
 tags:
   - laotian
   - nuns
-  - bhikkhuni
   - gender
 year: 2010
 journal: bsr

_content/monographs/sapiens_harari-y.md

@@ -7,6 +7,7 @@ olid: OL28326205M
 course: world
 tags:
   - evolution
+  - past
   - power
 year: 2011
 publisher: "Kinneret Zmora-Bitan Dvir"

_data/content.yml

@@ -25,3 +25,4 @@ ccmm: 1.12 # category match multiplier
 scmm: 1.08 # similar category multiplier
 fcmm: 1.2 # featured content multiplier
 tsmm: 0.5 # two star content multiplier
+ctmm: 0.5 # downweight canon book tag matches

_plugins/similar_content.rb

@@ -8,13 +8,15 @@ class SimilarContentFooterTag < Liquid::Tag
     @@similar_cats = nil
     @@content = nil
     @@content_for_tag = nil
+    @@canon_tags = nil
 
     def dataInit(v)
       puts "Prefetching data for similar_content footer..."
       @@config = v["site.data.content"]
       @@parents_for_tag = {}
       @@similar_cats = {}
       @@content = []
+      @@canon_tags = Set.new
       @@content_for_tag = Hash.new { |h, k| h[k] = Set.new }
       for cofefe in v["site.categories"]
         cat = cofefe.to_liquid.to_h
@@ -23,6 +25,9 @@ def dataInit(v)
       for cofefe in v["site.tags"]
         tag = cofefe.to_liquid.to_h
         @@parents_for_tag[tag["slug"]] = tag["parents"]
+        if tag["is_canon"]
+            @@canon_tags << tag["slug"]
+        end
       end
       for cofefe in v["site.content"]
         c = cofefe.to_liquid.to_h
@@ -133,16 +138,21 @@ def render(v)
             end
             if candidate["tags"]&.size&.nonzero? then
                 denom += candidate["tags"].size * @@config["tdms"]
-                for t in include_content["tags"]
-                    if candidate["tags"].include? t then
-                        score += @@config["ttms"]
+                for t in candidate["tags"]
+                    ttms = @@config["ttms"]
+                    if @@canon_tags.include? t and candidate["category"] == "canon"
+                        ttms *= @@config["ctmm"]
+                        denom -= @@config["tdms"] * (1.0 - @@config["ctmm"])
+                    end
+                    if include_content["tags"].include? t then
+                        score += ttms
                     else
                         if @@parents_for_tag[t] then
                             for p in @@parents_for_tag[t]
-                                if candidate["tags"].include? p then
+                                if include_content["tags"].include? p then
                                     score += @@config["tpms"]
                                     break
-                                elsif candidate["course"] == p then
+                                elsif include_content["course"] == p then
                                     score += @@config["tpms"]
                                     break
                                 end

_tags/canonical-poetry.md

@@ -2,7 +2,6 @@
 title: "Poetry of the Pāḷi Canon"
 status: unpublished
 parents: [sutta]
-is_canon: true # changes the layout slightly
 ---
 
 Mostly for the canonical poetry collections of the Kn, but also for the miscellaneous poetry found throughout The Canon.

_tags/rebirth-stories.md

@@ -2,7 +2,6 @@
 title: "Canonical Rebirth Stories"
 status: unpublished
 parents: [indian, sutta]
-is_canon: true # changes the layout slightly
 ---
 
 Rebirth stories in the Pāḷi Canon.

_tests/similar_content.md

@@ -62,7 +62,7 @@ partial_cases:
     - [mindfulness-in-plain-english_gunaratana, how-to-meditate_yuttadhammo]
 ---
 
-A series of test cases to measure the quality of the content recommendations algorithm.
+A series of hand-labeled examples to estimate the quality of the content recommendations algorithm.
 
 The passing rate at the bottom is a general indicator of the algo's recall rate. It's not meant to be 100%.
 
@@ -76,7 +76,7 @@ The passing rate at the bottom is a general indicator of the algo's recall rate.
 
 | Test Name | Status  |  Notes |
 |-----------|---------|--------|{% for test in cases %}{% assign fc = site.content | find: "slug", test[0] %}{% assign sc = site.content | find: "slug", test[1] %}{% capture cont_req %}{% assign include_content = fc %}{% similar_content %}{% endcapture %}
-| "[{{ fc.title | split: ':' | first }}]({{ fc.url }})" should recommend "[{{ sc.title | split: ':' | first }}]({{ sc.url }})" <details><code>{{ cont_req | strip_html | strip_newlines }}</code></details> | {% if cont_req contains test[1] %}{% assign succs = succs | plus: 1 %}Pass ✅{% else %}{% assign fails = fails | plus: 1 %}FAIL ❌{% endif %}  | of {% assign c = cont_req | split: "</li>" | size | minus: 2 %}{% assign simcount = simcount | plus: c %}{{ c }}  |{% endfor %}
+| "[{{ fc.title | split: ':' | first }}]({{ fc.url }})" should recommend "[{{ sc.title | split: ':' | first }}]({{ sc.url }})" <details><code>{{ cont_req | strip_html | strip_newlines }}</code></details> | {% if cont_req contains test[1] %}{% assign succs = succs | plus: 1 %}Included{% else %}{% assign fails = fails | plus: 1 %}Not included{% endif %}  | of {% assign c = cont_req | split: "</li>" | size | minus: 2 %}{% assign simcount = simcount | plus: c %}{{ c }}  |{% endfor %}
 |-----|------|----|
-| Totals: | {{ succs }} Pass and {{ fails }} Failed | {{ simcount }} total recommendations |
+| Totals: | {{ succs }} included out of {% assign total = fails | plus: succs %}{{total}} cases ({{ succs | times: 100 | divided_by: total }}%) | {{ simcount }} total recommendations |