HTTPArchive
diff --git a/‎sql/2024/fonts/README.md‎
Lines changed: 47 additions & 13 deletions b/‎sql/2024/fonts/README.md‎
Lines changed: 47 additions & 13 deletions
diff --git a/‎sql/2024/fonts/common.sql‎
Lines changed: 151 additions & 0 deletions b/‎sql/2024/fonts/common.sql‎
Lines changed: 151 additions & 0 deletions
diff --git a/‎sql/2024/fonts/design/fonts_designer.sql‎
Lines changed: 52 additions & 0 deletions b/‎sql/2024/fonts/design/fonts_designer.sql‎
Lines changed: 52 additions & 0 deletions
diff --git a/‎sql/2024/fonts/design/fonts_family_by_foundry.sql‎
Lines changed: 42 additions & 0 deletions b/‎sql/2024/fonts/design/fonts_family_by_foundry.sql‎
Lines changed: 42 additions & 0 deletions
diff --git a/‎sql/2024/fonts/design/fonts_family_by_script.sql‎
Lines changed: 46 additions & 0 deletions b/‎sql/2024/fonts/design/fonts_family_by_script.sql‎
Lines changed: 46 additions & 0 deletions
@@ -1,20 +1,54 @@
 # 2024 Fonts queries
 
-<!--
-  This directory contains all of the 2024 Fonts chapter queries.
+## Resources
 
-  Each query should have a corresponding `metric_name.sql` file.
-  Note that readers are linked to this directory, so try to make the SQL file names descriptive for easy browsing.
+* 📄 [Planning document]
+* 📊 [Results sheet]
+* 📝 [Chapter content]
 
-  Analysts: if helpful, you can use this README to give additional info about the queries.
--->
+## Structure
 
-## Resources
+The queries are split by the section where they are used:
+
+* `design/` is about foundries and families,
+* `development/` is about tools and technologies, and
+* `performance/` is about hosting and serving.
+
+Each file name starts with one of the following prefixes indicating the primary
+subject of the corresponding analysis:
+
+* `fonts_` is about font files,
+* `pages_` is about HTML pages,
+* `scripts_` is about JavaScript scripts, and
+* `styles_` is about CSS style sheets.
+
+The prefix is followed by the property studied given in singular, potentially
+extended one or several suffixes narrowing down the scope, as in
+`fonts_size_by_table.sql` and `pages_link_relation.sql`.
+
+## Contents
+
+Each query starts with a preamble indicating the section, question, and
+normalization type:
+
+```sql
+-- Section: Performance
+-- Question: What is the distribution of the file size broken down by table?
+-- Normalization: Pages
+```
+
+Many queries rely on temporary functions for convenience and clarity. The
+functions appear in several queries are extracted into a common file:
+`common.sql`. Whenever any of the functions defined in `common.sql` is used by a
+query, the query has the following line at the top:
+
+```sql
+-- INCLUDE https://github.com/HTTPArchive/almanac.httparchive.org/blob/main/sql/2024/fonts/common.sql
+```
 
-- [📄 Planning doc][~google-doc]
-- [📊 Results sheet][~google-sheets]
-- [📝 Markdown file][~chapter-markdown]
+It signalizes that, prior to executing the query, `common.sql` has to be
+inlined.
 
-[~google-doc]: https://docs.google.com/document/d/1ljEHbDvXComXnW5s_EXZ0nM3_JCLnYr28Xrcf0YYtP8/edit#heading=h.vp0ukb2pxxzp
-[~google-sheets]: https://docs.google.com/spreadsheets/d/1EkdvJ8e0B9Rr42evC2Ds5Ekwq6gF9oLBW0BA5cmSUT4/edit#gid=1778117656
-[~chapter-markdown]: https://github.com/HTTPArchive/almanac.httparchive.org/tree/main/src/content/en/2024/fonts.md
+[Planning document]: https://docs.google.com/document/d/1ljEHbDvXComXnW5s_EXZ0nM3_JCLnYr28Xrcf0YYtP8/edit
+[Results sheet]: https://docs.google.com/spreadsheets/d/1EkdvJ8e0B9Rr42evC2Ds5Ekwq6gF9oLBW0BA5cmSUT4/edit
+[Chapter content]: https://github.com/HTTPArchive/almanac.httparchive.org/tree/main/src/content/en/2024/fonts.md
@@ -0,0 +1,151 @@
+-- Normalize a family name. Used in FAMILY_INNER.
+CREATE TEMPORARY FUNCTION FAMILY_INNER_INNER(name STRING) AS (
+  CASE
+    WHEN REGEXP_CONTAINS(name, r'(?i)font\s?awesome') THEN 'Font Awesome'
+    ELSE IF(LENGTH(TRIM(name)) < 3, NULL, NULLIF(TRIM(name), ''))
+  END
+);
+
+-- Normalize a family name. Used in FAMILY.
+CREATE TEMPORARY FUNCTION FAMILY_INNER(name STRING) AS (
+  FAMILY_INNER_INNER(
+    REGEXP_REPLACE(
+      name,
+      r'(?i)([\s-]?(black|bold|book|cond(ensed)?|demi|ex(tra)?|heavy|italic|light|medium|narrow|regular|semi|thin|ultra|wide|\d00|\d+pt))+$',
+      ''
+    )
+  )
+);
+
+-- Extract the family name from a payload.
+CREATE TEMPORARY FUNCTION FAMILY(payload STRING) AS (
+  FAMILY_INNER(
+    COALESCE(
+      JSON_EXTRACT_SCALAR(payload, '$._font_details.names[16]'),
+      JSON_EXTRACT_SCALAR(payload, '$._font_details.names[1]')
+    )
+  )
+);
+
+-- Extract the file format from an extension and a MIME type.
+CREATE TEMPORARY FUNCTION FILE_FORMAT(extension STRING, type STRING) AS (
+  LOWER(IFNULL(REGEXP_EXTRACT(type, '/(?:x-)?(?:font-)?(.*)'), extension))
+);
+
+-- Normalize a foundry name. Used in FOUNDRY.
+CREATE TEMPORARY FUNCTION FOUNDRY_INNER(name STRING) AS (
+  CASE UPPER(name)
+    WHEN 'ADBO' THEN 'ADBE'
+    WHEN 'PFED' THEN 'AWSM'
+    ELSE NULLIF(TRIM(REGEXP_REPLACE(name, r'[[:cntrl:]]+', '')), '')
+  END
+);
+
+-- Extract the foundry name from a payload.
+CREATE TEMPORARY FUNCTION FOUNDRY(payload STRING) AS (
+  FOUNDRY_INNER(JSON_EXTRACT_SCALAR(payload, '$._font_details.OS2.achVendID'))
+);
+
+-- Infer scripts from codepoints. Used in SCRIPTS.
+CREATE TEMPORARY FUNCTION SCRIPTS_INNER(codepoints ARRAY<STRING>)
+RETURNS ARRAY<STRING>
+LANGUAGE js
+OPTIONS (library = ["gs://httparchive/lib/text-utils.js"])
+AS r"""
+if (codepoints && codepoints.length) {
+  return detectWritingScript(codepoints.map((character) => parseInt(character, 10)), 0.05);
+} else {
+  return [];
+}
+""";
+
+-- Infer scripts from a payload.
+CREATE TEMPORARY FUNCTION SCRIPTS(payload STRING) AS (
+  SCRIPTS_INNER(JSON_EXTRACT_STRING_ARRAY(payload, '$._font_details.cmap.codepoints'))
+);
+
+-- Infer the service from a URL.
+CREATE TEMPORARY FUNCTION SERVICE(url STRING) AS (
+  CASE
+    WHEN REGEXP_CONTAINS(url, r'(fonts|use)\.typekit\.(net|com)') THEN 'Adobe'
+    WHEN REGEXP_CONTAINS(url, r'cloud\.typenetwork\.com') THEN 'typenetwork.com'
+    WHEN REGEXP_CONTAINS(url, r'cloud\.typography\.com') THEN 'typography.com'
+    WHEN REGEXP_CONTAINS(url, r'cloud\.webtype\.com') THEN 'webtype.com'
+    WHEN REGEXP_CONTAINS(url, r'f\.fontdeck\.com') THEN 'fontdeck.com'
+    WHEN REGEXP_CONTAINS(url, r'fast\.fonts\.(com|net)\/(jsapi|cssapi)') THEN 'fonts.com'
+    WHEN REGEXP_CONTAINS(url, r'fnt\.webink\.com') THEN 'webink.com'
+    WHEN REGEXP_CONTAINS(url, r'fontawesome\.com') THEN 'fontawesome.com'
+    WHEN REGEXP_CONTAINS(url, r'fonts\.(gstatic|googleapis)\.com|themes.googleusercontent.com/static/fonts|ssl.gstatic.com/fonts') THEN 'Google'
+    WHEN REGEXP_CONTAINS(url, r'fonts\.typonine\.com') THEN 'typonine.com'
+    WHEN REGEXP_CONTAINS(url, r'fonts\.typotheque\.com') THEN 'typotheque.com'
+    WHEN REGEXP_CONTAINS(url, r'kernest\.com') THEN 'kernest.com'
+    WHEN REGEXP_CONTAINS(url, r'typefront\.com') THEN 'typefront.com'
+    WHEN REGEXP_CONTAINS(url, r'typesquare\.com') THEN 'typesquare.com'
+    WHEN REGEXP_CONTAINS(url, r'use\.edgefonts\.net|webfonts\.creativecloud\.com') THEN 'edgefonts.net'
+    WHEN REGEXP_CONTAINS(url, r'webfont\.fontplus\.jp') THEN 'fontplus.jp'
+    WHEN REGEXP_CONTAINS(url, r'webfonts\.fontslive\.com') THEN 'fontslive.com'
+    WHEN REGEXP_CONTAINS(url, r'webfonts\.fontstand\.com') THEN 'fontstand.com'
+    WHEN REGEXP_CONTAINS(url, r'webfonts\.justanotherfoundry\.com') THEN 'justanotherfoundry.com'
+    ELSE 'self-hosted'
+  END
+);
+
+-- Extract the color formats from a formats payload and remove spurious entries
+-- via a table-sizes payload.
+--
+-- When nonempty, it is expected that
+--
+-- * `CBDT` is larger than 2 + 2 bytes,
+-- * `COLR` is larger than 2 + 2 + 4 + 4 + 2 (+ 4 + 4 + 4 + 4 + 4) bytes,
+-- * `SVG ` is larger than 2 + 4 + 4 + 2 bytes, and
+-- * `sbix` is larger than 2 + 2 + 4 + 4 bytes.
+--
+-- For simplicity, the threshold is set to 50 bytes.
+CREATE TEMPORARY FUNCTION COLOR_FORMATS_INNER(jsonFormats STRING, jsonSizes STRING)
+RETURNS ARRAY<STRING>
+LANGUAGE js AS '''
+try {
+  const formats = JSON.parse(jsonFormats);
+  const sizes = JSON.parse(jsonSizes);
+  return formats.filter((format) => {
+    const table = `${format}    `.slice(0, 4);
+    return sizes[table] > 50;
+  });
+} catch (e) {
+  return [];
+}
+''';
+
+-- Extract the color formats from a payload.
+CREATE TEMPORARY FUNCTION COLOR_FORMATS(payload STRING) AS (
+  COLOR_FORMATS_INNER(
+    JSON_EXTRACT(payload, '$._font_details.color.formats'),
+    JSON_EXTRACT(payload, '$._font_details.table_sizes')
+  )
+);
+
+-- Check if the font is a color font given its payload.
+CREATE TEMPORARY FUNCTION IS_COLOR(payload STRING) AS (
+  ARRAY_LENGTH(COLOR_FORMATS(payload)) > 0
+);
+
+-- Check if the font was successfully parsed given its payload.
+CREATE TEMPORARY FUNCTION IS_PARSED(payload STRING) AS (
+  JSON_EXTRACT(payload, '$._font_details.table_sizes') IS NOT NULL
+);
+
+-- Check if the font is a variable font given its payload.
+CREATE TEMPORARY FUNCTION IS_VARIABLE(payload STRING) AS (
+  REGEXP_CONTAINS(
+    JSON_EXTRACT(payload, '$._font_details.table_sizes'),
+    '(?i)gvar|CFF2'
+  )
+);
+
+-- Extract the variable formats from a payload.
+CREATE TEMPORARY FUNCTION VARIABLE_FORMATS(payload STRING) AS (
+  REGEXP_EXTRACT_ALL(
+    JSON_EXTRACT(payload, '$._font_details.table_sizes'),
+    '(?i)glyf|CFF2'
+  )
+);
@@ -0,0 +1,52 @@
+-- Section: Design
+-- Question: Which designers are popular?
+-- Normalization: Pages
+
+-- INCLUDE https://github.com/HTTPArchive/almanac.httparchive.org/blob/main/sql/2024/fonts/common.sql
+
+WITH
+designers AS (
+  SELECT
+    client,
+    NULLIF(TRIM(JSON_EXTRACT_SCALAR(payload, '$._font_details.names[9]')), '') AS designer,
+    COUNT(DISTINCT page) AS count,
+    ROW_NUMBER() OVER (PARTITION BY client ORDER BY COUNT(DISTINCT page) DESC) AS rank
+  FROM
+    `httparchive.all.requests`
+  WHERE
+    date = '2024-07-01' AND
+    type = 'font' AND
+    is_root_page AND
+    IS_PARSED(payload)
+  GROUP BY
+    client,
+    designer
+  QUALIFY
+    rank <= 100
+),
+pages AS (
+  SELECT
+    client,
+    COUNT(DISTINCT page) AS total
+  FROM
+    `httparchive.all.requests`
+  WHERE
+    date = '2024-07-01' AND
+    is_root_page
+  GROUP BY
+    client
+)
+
+SELECT
+  client,
+  designer,
+  count,
+  total,
+  count / total AS proportion
+FROM
+  designers
+JOIN
+  pages USING (client)
+ORDER BY
+  client,
+  proportion DESC
@@ -0,0 +1,42 @@
+-- Section: Design
+-- Question: Which families are used broken down by foundry?
+-- Normalization: Requests (parsed only)
+
+-- INCLUDE https://github.com/HTTPArchive/almanac.httparchive.org/blob/main/sql/2024/fonts/common.sql
+
+WITH
+requests AS (
+  SELECT
+    client,
+    FOUNDRY(payload) AS foundry,
+    FAMILY(payload) AS family,
+    COUNT(0) OVER (PARTITION BY client) AS total
+  FROM
+    `httparchive.all.requests`
+  WHERE
+    date = '2024-07-01' AND
+    type = 'font' AND
+    IS_PARSED(payload) AND
+    is_root_page
+)
+
+SELECT
+  client,
+  foundry,
+  family,
+  COUNT(0) AS count,
+  total,
+  COUNT(0) / total AS proportion,
+  ROW_NUMBER() OVER (PARTITION BY client ORDER BY COUNT(0) DESC) AS rank
+FROM
+  requests
+GROUP BY
+  client,
+  foundry,
+  family,
+  total
+QUALIFY
+  rank <= 100
+ORDER BY
+  client,
+  proportion DESC
@@ -0,0 +1,46 @@
+-- Section: Design
+-- Question: Which families are used broken down by script?
+-- Normalization: Requests (parsed only)
+
+-- INCLUDE https://github.com/HTTPArchive/almanac.httparchive.org/blob/main/sql/2024/fonts/common.sql
+
+WITH
+requests AS (
+  SELECT
+    client,
+    SCRIPTS(payload) AS scripts,
+    FAMILY(payload) AS family,
+    COUNT(0) OVER (PARTITION BY client) AS total
+  FROM
+    `httparchive.all.requests`
+  WHERE
+    date = '2024-07-01' AND
+    type = 'font' AND
+    is_root_page AND
+    IS_PARSED(payload)
+)
+
+SELECT
+  client,
+  script,
+  family,
+  COUNT(0) AS count,
+  total,
+  COUNT(0) / total AS proportion,
+  ROW_NUMBER() OVER (PARTITION BY client, script ORDER BY COUNT(0) DESC) AS rank
+FROM
+  requests,
+  UNNEST(scripts) AS script
+WHERE
+  family != 'Adobe Blank'
+GROUP BY
+  client,
+  script,
+  family,
+  requests.total
+QUALIFY
+  rank <= 10
+ORDER BY
+  client,
+  script,
+  proportion DESC