present per document metadata

author: Ralph Amissah <ralph.amissah@gmail.com> 2019-12-13 09:09:24 -0500
committer: Ralph Amissah <ralph.amissah@gmail.com> 2020-01-13 16:06:43 -0500
commit: 0c4b3cb3707c3b16cd171620427e651d71182813 (patch)
tree: 2027a8f9449eab7b2b878ee9f1554a417563b73f /org/out_metadata.org
parent: harvest topics, name hashtag level names (diff)
1 files changed, 486 insertions, 0 deletions
diff --git a/org/out_metadata.org b/org/out_metadata.org
new file mode 100644
index 0000000..f48d6ac
--- /dev/null
+++ b/org/out_metadata.org
@@ -0,0 +1,486 @@
+-*- mode: org -*-
+#+TITLE:       spine (doc_manifest) metadata (document harvest)
+#+DESCRIPTION: documents - structuring, publishing in multiple formats & search
+#+FILETAGS:    :spine:output:metadata:
+#+AUTHOR:      Ralph Amissah
+#+EMAIL:       [[mailto:ralph.amissah@gmail.com][ralph.amissah@gmail.com]]
+#+COPYRIGHT:   Copyright (C) 2015 - 2019 Ralph Amissah
+#+LANGUAGE:    en
+#+STARTUP:     indent content hideblocks hidestars
+#+OPTIONS:     H:3 num:nil toc:t \n:nil @:t ::t |:t ^:nil _:nil -:t f:t *:t <:t
+#+OPTIONS:     TeX:t LaTeX:t skip:nil d:nil todo:t pri:nil tags:not-in-toc
+#+OPTIONS:     author:nil email:nil creator:nil timestamp:nil
+#+PROPERTY:    header-args :padline no :exports code :cache no :noweb yes
+#+EXPORT_SELECT_TAGS:  export
+#+EXPORT_EXCLUDE_TAGS: noexport
+#+TAGS: assert(a) class(c) debug(d) mixin(m) spine(s) tangle(T) template(t) WEB(W) noexport(n)
+
+* document harvest _summary_         :module:spine:metadoc_show_summary:
+
+** _module template_ metadoc document metadata
+
+#+BEGIN_SRC d :tangle "../src/spine/io_out/metadata.d"
+module spine.io_out.metadata;
+template outputMetadata() {
+  void outputMetadata(T)( T  doc_matters) @safe {
+    <<output_imports>>
+<<harvested_html_themes>>
+<<harvested_topics_html_head_1>>
+<<harvested_html_head>>
+<<harvested_topics_html_head_2>>
+<<harvested_topics_html_head_theme>>
+    <<output_metadata_write>>
+    <<output_metadata>>
+  }
+}
+#+END_SRC
+
+** imports
+
+#+name: output_imports
+#+BEGIN_SRC d
+import std.file;
+import std.format;
+import spine.io_out;
+mixin InternalMarkup;
+string[] metadata_;
+#+END_SRC
+
+** metadata
+
+#+name: output_metadata
+#+BEGIN_SRC d
+static auto mkup = InlineMarkup();
+if (doc_matters.opt.action.debug_do) {
+  writeln(doc_matters.src.filename_base);
+  writeln("Title:       ", doc_matters.conf_make_meta.meta.title_full);
+  writeln("  Author:    ", doc_matters.conf_make_meta.meta.creator_author);
+  writeln("  Published: ", doc_matters.conf_make_meta.meta.date_published);
+  writeln("  Copyright: ", doc_matters.conf_make_meta.meta.rights_copyright);
+  writeln("  License:   ", doc_matters.conf_make_meta.meta.rights_license);
+  if (doc_matters.conf_make_meta.meta.classify_topic_register_arr.length > 0) {
+    foreach (topic; doc_matters.conf_make_meta.meta.classify_topic_register_arr.sort!("toUpper(a) < toUpper(b)", SwapStrategy.unstable)) {
+      string[] subject_tree = topic.split(mkup.sep);
+      if (subject_tree.length > 0) { writeln("  ",         subject_tree[0]); }
+      if (subject_tree.length > 1) { writeln("    ",       subject_tree[1]); }
+      if (subject_tree.length > 2) { writeln("      ",     subject_tree[2]); }
+      if (subject_tree.length > 3) { writeln("        ",   subject_tree[3]); }
+      if (subject_tree.length > 4) { writeln("          ", subject_tree[4]); }
+    }
+  }
+}
+auto pth_html = spinePathsHTML!()(doc_matters.output_path, doc_matters.src.language);
+auto pth_epub = spinePathsEPUB!()(doc_matters.output_path, doc_matters.src.language);
+auto pth_pod =  spinePathsPods!()(doc_matters);
+metadata_ ~= format(q"┃<body lang="en" xml:lang="en">
+<a name="top" id="top"></a>
+<a name="up" id="up"></a>
+<a name="start" id="start"></a>
+┃");
+if (doc_matters.opt.action.harvest_link) {
+  metadata_ ~= format(q"┃<p>[<a href="../../index.html">&nbsp;HOME&nbsp;</a>] Metadata Harvest
+   [<a href="../../authors.html">&nbsp;Authors&nbsp;</a>]
+   [<a href="../../topics.html">&nbsp;Topics&nbsp;</a>]</p>
+<hr />
+┃");
+} else {
+  metadata_ ~= format(q"┃<p>[<a href="../../index.html">&nbsp;HOME&nbsp;</a>]
+<hr />
+┃");
+}
+if (!(doc_matters.conf_make_meta.meta.title_full.empty)) {
+  metadata_ ~= "<p class=\"lev0\">Title: <b><a href=\"" ~ doc_matters.src.filename_base ~ "/toc.html\">" ~ doc_matters.conf_make_meta.meta.title_full       ~ "</a></b></p>";
+} else if (doc_matters.opt.action.debug_do || doc_matters.opt.action.very_verbose) {
+  writeln("ERROR no Title information provided in document header ", doc_matters.src.filename_base);
+}
+if (!(doc_matters.conf_make_meta.meta.creator_author.empty)) {
+  if (doc_matters.opt.action.harvest_link) {
+    metadata_ ~= "<p class=\"lev1\">Author: <b><a href=\"../../authors.html#" ~ doc_matters.conf_make_meta.meta.creator_author_surname.translate([' ' : "_"]) ~ "\">"
+                                               ~ doc_matters.conf_make_meta.meta.creator_author   ~ "</a></b></p>";
+  } else {
+    metadata_ ~= "<p class=\"lev1\">Author: <b>"
+                                               ~ doc_matters.conf_make_meta.meta.creator_author   ~ "</b></p>";
+  }
+} else if (doc_matters.opt.action.debug_do || doc_matters.opt.action.very_verbose) {
+  writeln("ERROR no Author information provided in document header ", doc_matters.src.filename_base);
+}
+metadata_ ~= "<p class=\"lev1\">Published: "   ~ doc_matters.conf_make_meta.meta.date_published   ~ "</p>";
+if (!(doc_matters.conf_make_meta.meta.rights_copyright.empty)) {
+  metadata_ ~= "<p class=\"lev1\">Copyright: "   ~ doc_matters.conf_make_meta.meta.rights_copyright ~ "</p>";
+} else if (doc_matters.opt.action.debug_do || doc_matters.opt.action.very_verbose) {
+  writeln("WARNING no Copyright information provided in document header ", doc_matters.src.filename_base);
+}
+if (!(doc_matters.conf_make_meta.meta.rights_license.empty)) {
+  metadata_ ~= "<p class=\"lev1\">License:   "   ~ doc_matters.conf_make_meta.meta.rights_license   ~ "</p>";
+} else if (doc_matters.opt.action.debug_do || doc_matters.opt.action.very_verbose) {
+  writeln("WARNING no License information provided in document header ", doc_matters.src.filename_base);
+}
+if (!(doc_matters.conf_make_meta.meta.notes_summary.empty)) {
+  metadata_ ~= "<hr /><p class=\"lev0\">Summary:</p><p class=\"lev1\">"   ~ doc_matters.conf_make_meta.meta.notes_summary   ~ "</p>";
+} else if (doc_matters.opt.action.debug_do) {
+  writeln("WARNING no summary of text provided in document header ", doc_matters.src.filename_base);
+}
+if (doc_matters.conf_make_meta.meta.classify_topic_register_arr.length > 0) {
+  metadata_ ~= "<hr /><p class=\"lev0\">Topics:</p>";
+  string[] _top = ["", "", "", "", ""];
+  foreach (topic; doc_matters.conf_make_meta.meta.classify_topic_register_arr.sort!("toUpper(a) < toUpper(b)", SwapStrategy.unstable)) {
+    string[] subject_tree = topic.split(mkup.sep);
+    if (subject_tree.length > 0) {
+      if (subject_tree[0] != _top[0]) {
+        _top[0] = subject_tree[0];
+        if (doc_matters.opt.action.harvest_link) {
+          metadata_ ~=
+            "<p class=\"lev1\"><a href=\"../../topics.html#"
+              ~ subject_tree[0].translate([' ' : "_"]) ~ "\">"
+              ~ subject_tree[0]
+              ~ "</a></p>";
+        } else {
+          metadata_ ~=
+            "<p class=\"lev1\">"  ~ subject_tree[0] ~ "</p>";
+        }
+      }
+      if (subject_tree.length > 1) {
+        if (subject_tree[1] != _top[1]) {
+          _top[1] = subject_tree[1];
+          _top[2] = ""; _top[3] = ""; _top[4] = "";
+          if (doc_matters.opt.action.harvest_link) {
+            metadata_ ~=
+              "<p class=\"lev2\"><a href=\"../../topics.html#"
+                ~ subject_tree[0].translate([' ' : "_"]) ~ "."
+                ~ subject_tree[1].translate([' ' : "_"]) ~ "\">"
+                ~ subject_tree[1]
+                ~ "</a></p>";
+          } else {
+            metadata_ ~=
+              "<p class=\"lev2\">"  ~ subject_tree[1] ~ "</p>";
+          }
+        }
+        if (subject_tree.length > 2) {
+          if (subject_tree[2] != _top[2]) {
+            _top[2] = subject_tree[2];
+            _top[3] = ""; _top[4] = "";
+            if (doc_matters.opt.action.harvest_link) {
+              metadata_ ~=
+                "<p class=\"lev3\"><a href=\"../../topics.html#"
+                  ~ subject_tree[0].translate([' ' : "_"]) ~ "."
+                  ~ subject_tree[1].translate([' ' : "_"]) ~ "."
+                  ~ subject_tree[2].translate([' ' : "_"]) ~ "\">"
+                  ~ subject_tree[2]
+                  ~ "</a></p>";
+            } else {
+              metadata_ ~=
+                "<p class=\"lev3\">"  ~ subject_tree[2] ~ "</p>";
+            }
+          }
+          if (subject_tree.length > 3) {
+            if (subject_tree[3] != _top[3]) {
+              _top[3] = subject_tree[3];
+              _top[4] = "";
+              if (doc_matters.opt.action.harvest_link) {
+                metadata_ ~=
+                  "<p class=\"lev4\"><a href=\"../../topics.html#"
+                    ~ subject_tree[0].translate([' ' : "_"]) ~ "."
+                    ~ subject_tree[1].translate([' ' : "_"]) ~ "."
+                    ~ subject_tree[2].translate([' ' : "_"]) ~ "."
+                    ~ subject_tree[3].translate([' ' : "_"]) ~ "\">"
+                    ~ subject_tree[3]
+                    ~ "</a></p>";
+              } else {
+                metadata_ ~=
+                  "<p class=\"lev4\">"  ~ subject_tree[3] ~ "</p>";
+              }
+            }
+            if (subject_tree.length > 4) {
+              if (subject_tree[4] != _top[4]) {
+                _top[4] = subject_tree[4];
+                if (doc_matters.opt.action.harvest_link) {
+                  metadata_ ~=
+                    "<p class=\"lev5\"><a href=\"../../topics.html#"
+                      ~ subject_tree[0].translate([' ' : "_"]) ~ "."
+                      ~ subject_tree[1].translate([' ' : "_"]) ~ "."
+                      ~ subject_tree[2].translate([' ' : "_"]) ~ "."
+                      ~ subject_tree[3].translate([' ' : "_"]) ~ "."
+                      ~ subject_tree[4].translate([' ' : "_"]) ~ "\">"
+                      ~ subject_tree[4]
+                      ~ "</a></p>";
+                } else {
+                  metadata_ ~=
+                    "<p class=\"lev5\">"  ~ subject_tree[4] ~ "</p>";
+                }
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+} else if (doc_matters.opt.action.debug_do) {
+  writeln("WARNING no topic_register classification of text provided in document header ", doc_matters.src.filename_base);
+}
+metadata_ ~= "<hr /><p class=\"lev1\">source: "      ~ doc_matters.src.filename_base ~ "</p>";
+metadata_ ~= "<p class=\"lev1\">markup source: [<a href=\"../../" ~ pth_pod.internal_base ~ "/" ~ doc_matters.src.filename_base ~ ".zip\">"
+             ~ "&nbsp;zipped&nbsp;pod&nbsp;</a>] "
+             ~ "[<a href=\"../../" ~ pth_pod.internal_base ~ "/" ~ doc_matters.src.filename_base ~ "/\">"
+             ~ "&nbsp;pod&nbsp;tree&nbsp;</a>]</p>";
+metadata_ ~= "<p class=\"lev1\">outputs: [<a href=\""    ~ doc_matters.src.filename_base ~ ".html\">"
+             ~ "&nbsp;html&nbsp;scroll&nbsp;</a>] "
+             ~ "[<a href=\""    ~ doc_matters.src.filename_base ~ "/toc.html\">"
+             ~ "&nbsp;html&nbsp;seg&nbsp;</a>]"
+             ~ "[<a href=\"../../" ~ pth_epub.internal_base ~ "/" ~ doc_matters.src.filename_base ~ "." ~ doc_matters.src.language ~ ".epub\">"
+             ~ "&nbsp;epub&nbsp;</a>]</p>";
+metadata_write_output(doc_matters, metadata_);
+#+END_SRC
+
+** metadata write output
+
+#+name: output_metadata_write
+#+BEGIN_SRC d
+void metadata_write_output(M)(M doc_matters, string[] metadata_) @trusted {
+  auto pth_html = spinePathsHTML!()(doc_matters.output_path, doc_matters.src.language);
+  try {
+    if (!exists(pth_html.base)) {
+      pth_html.base.mkdirRecurse;
+    }
+    auto f = File(pth_html.fn_scroll("metadata." ~ doc_matters.src.filename), "w");
+    foreach (o; metadata_) {
+      f.writeln(o);
+    }
+  } catch (ErrnoException ex) {
+    // Handle error
+  }
+  if (!(doc_matters.opt.action.quiet)) {
+    writeln(" ", pth_html.fn_scroll("metadata." ~ doc_matters.src.filename));
+  }
+}
+#+END_SRC
+
+*** themes
+**** head
+
+#+NAME: harvested_html_themes
+#+BEGIN_SRC d
+string theme_dark_0 = format(q"┃
+  body {
+    color                    : #CCCCCC;
+    background               : #000000;
+    background-color         : #000000;
+  }
+  a:link {
+    color                    : #FFFFFF;
+    text-decoration          : none;
+  }
+  a:visited {
+    color                    : #999999;
+    text-decoration          : none;
+  }
+  a:hover {
+    color                    : #000000;
+    background-color         : #555555;
+  }
+  a:hover img {
+    background-color         : #000000;
+  }
+  a:active {
+    color                    : #888888;
+    text-decoration          : underline;
+  }
+┃");
+string theme_light_0 = format(q"┃
+  body {
+    color                    : #000000;
+    background               : #FFFFFF;
+    background-color         : #FFFFFF;
+  }
+  a:link {
+    color                    : #003399;
+    text-decoration          : none;
+  }
+  a:visited {
+    color                    : #003399;
+    text-decoration          : none;
+  }
+  a:hover {
+    color                    : #000000;
+    background-color         : #f9f9aa;
+  }
+  a:hover img {
+    background-color         : #FFFFFF;
+  }
+  a:active {
+    color                    : #003399;
+    text-decoration          : underline;
+  }
+┃");
+#+END_SRC
+
+**** levels
+
+#+NAME: harvested_html_themes
+#+BEGIN_SRC d
+string theme_dark_1 = format(q"┃
+  h1 {
+    color                    : #FFFFFF;
+    background               : #000000;
+  }
+  p.letter {
+    color                    : #FFFFFF;
+    background               : #333333;
+  }
+┃");
+string theme_light_1 = format(q"┃
+  h1 {
+    color                    : #FFFFFF;
+    background               : #000088;
+  }
+  p.letter {
+    color                    : #FFFFFF;
+    background               : #880000;
+  }
+┃");
+#+END_SRC
+
+*** harvested topics html head
+
+#+NAME: harvested_topics_html_head_1
+#+BEGIN_SRC d
+      metadata_ ~= format(q"┃<!DOCTYPE html>
+<html>
+<head>
+<meta charset="utf-8">
+<title>Metadata Harvest - Topics</title>
+#+END_SRC
+
+#+NAME: harvested_topics_html_head_2
+#+BEGIN_SRC d
+</head>
+#+END_SRC
+
+*** harvested html head
+
+#+NAME: harvested_html_head
+#+BEGIN_SRC d
+<meta http-equiv="Content-Type" content="text/html;charset=utf-8" />
+<meta name="dc.title" content= "metadata harvest, Authors & Topics - information Structuring Universe, Structured information Serialised Units" />
+<meta name="dc.subject" content= "document structuring, ebook, publishing, PDF, LaTeX, XML, ODF, SQL, postgresql, sqlite, electronic book, electronic publishing, electronic document, electronic citation, data structure, citation systems, granular search, digital library" />
+<meta name="generator" content="spine" />
+<link rel="generator" href="http://sisudoc.org" />
+<link href="./css/harvest.css" rel="stylesheet">
+<style TYPE="text/css">
+/* spine harvest css default stylesheet */%s
+  .norm, .bold {
+    line-height              : 150%%;
+    margin-left              : 1em;
+    margin-right             : 2em;
+    margin-top               : 10px;
+    margin-bottom            : 0px;
+    text-indent              : 0mm;
+  }
+  p, h0, h1, h2, h3, h4, h5, h6, h7 {
+    display                  : block;
+    font-family              : verdana, arial, georgia, tahoma, sans-serif, helvetica, times, roman;
+    font-size                : 100%%;
+    font-weight              : normal;
+    line-height              : 150%%;
+    /* text-align            : justify; */
+    margin-left              : 1em;
+    text-indent              : 0mm;
+    margin-top               : 2px;
+    margin-bottom            : 2px;
+    margin-right             : 6px;
+    text-align               : left;
+  }
+  h0, h1, h2, h3, h4, h5, h6, h7 { text-shadow: .2em .2em .3em #999999; }
+  h1 {
+    font-size                : 120%%;
+    font-weight              : bold;
+    color                    : #FFFFFF;
+    background               : #000088;
+    margin-left              : 0em;
+  }
+  p.work {
+    font-size                : 80%%;
+    margin-left              : 5em;
+    margin-top               : 0px;
+    margin-bottom            : 0px;
+    margin-right             : 6px;
+    text-align               : left;
+  }
+  p.author {
+    font-size                : 100%%;
+    margin-left              : 2em;
+    margin-top               : 0px;
+    margin-bottom            : 0px;
+    margin-right             : 6px;
+    text-align               : left;
+  }
+  p.publication {
+    font-size                : 80%%;
+    margin-left              : 4em;
+    margin-top               : 0px;
+    margin-bottom            : 0px;
+    margin-right             : 6px;
+    text-align               : left;
+  }
+  p.letter {
+    font-weight              : bold;
+    font-size                : 60%%;
+    margin-left              : 1em;
+    margin-top               : 0px;
+    margin-bottom            : 0px;
+    margin-right             : 6px;
+    text-align               : left;
+  }
+  p.lev0 {
+    font-size                : 120%%;
+    margin-left              : 1em;
+  }
+  p.lev1 {
+    font-size                : 110%%;
+    margin-left              : 2em;
+  }
+  p.lev2 {
+    font-size                : 100%%;
+    margin-left              : 3em;
+  }
+  p.lev3 {
+    font-size                : 90%%;
+    margin-left              : 4em;
+  }
+  p.lev4 {
+    font-size                : 80%%;
+    margin-left              : 5em;
+  }
+  p.lev5 {
+    font-size                : 80%%;
+    margin-left              : 6em;
+  }%s
+</style>
+<link rel="shortcut icon" href="../_sisu/image/rb7.ico" />
+#+END_SRC
+
+*** harvested html bottom
+
+#+NAME: harvested_html_bottom
+#+BEGIN_SRC d
+      ~= format(q"┃
+<hr />
+<a name="bottom" id="bottom"></a>
+<a name="down" id="down"></a>
+<a name="end" id="end"></a>
+<a name="finish" id="finish"></a>
+<a name="stop" id="stop"></a>
+<a name="credits"></a>
+</body>
+</html>
+┃") ~ "\n";
+#+END_SRC
+
+*** harvested topics html head theme insert
+
+#+NAME: harvested_topics_html_head_theme
+#+BEGIN_SRC d
+┃",
+  doc_matters.opt.action.css_theme_default ? theme_light_0 : theme_dark_0,
+  doc_matters.opt.action.css_theme_default ? theme_light_1 : theme_dark_1,
+) ~ "\n";
+#+END_SRC
author	Ralph Amissah <ralph.amissah@gmail.com>	2019-12-13 09:09:24 -0500
committer	Ralph Amissah <ralph.amissah@gmail.com>	2020-01-13 16:06:43 -0500
commit	0c4b3cb3707c3b16cd171620427e651d71182813 (patch)
tree	2027a8f9449eab7b2b878ee9f1554a417563b73f /org/out_metadata.org
parent	harvest topics, name hashtag level names (diff)