harvest reorganized

author: Ralph Amissah <ralph.amissah@gmail.com> 2019-06-20 13:45:58 -0400
committer: Ralph Amissah <ralph.amissah@gmail.com> 2019-10-17 19:07:20 -0400
commit: 2dd7d2083c23fe9e79a984c534c2283fc4d7e581 (patch)
tree: 22deaa43f6888ea6c324a27b5fd4ec9b3546fb63 /src/doc_reform/meta/metadoc_harvests_topics.d
parent: harvest html output (diff)
1 files changed, 408 insertions, 0 deletions
diff --git a/src/doc_reform/meta/metadoc_harvests_topics.d b/src/doc_reform/meta/metadoc_harvests_topics.d
new file mode 100644
index 0000000..f0aed57
--- /dev/null
+++ b/src/doc_reform/meta/metadoc_harvests_topics.d
@@ -0,0 +1,408 @@
+module doc_reform.meta.metadoc_harvests_topics;
+  import
+    std.algorithm,
+    std.array,
+    std.exception,
+    std.regex,
+    std.stdio,
+    std.string,
+    std.conv : to;
+  import
+    doc_reform.meta.defaults,
+    doc_reform.meta.rgx;
+  mixin DocReformHarvest;
+  mixin InternalMarkup;
+  mixin DocReformRgxInit;
+template DocReformMetaDocHarvestsTopics() {
+  auto mkup = InlineMarkup();
+  void DocReformMetaDocHarvestsTopics(H,O)(
+    H  hvst,
+    O  _opt_action,
+  ) {
+      auto min_repeat_number = 42;
+      string[] _document_topic_register;
+      string[] _topic_register;
+      string[] _sub_topic_register;
+      string[] topics = [];
+      string _auth = "";
+      foreach(k, doc_harvest; hvst.harvests) {
+        _topic_register = [];
+        foreach(topic; doc_harvest.topic_register_arr.sort) {
+          _sub_topic_register = [];
+          string _spaces;
+          string[] subject_tree = topic.split(mkup.sep);
+          switch (subject_tree.length) {
+          case 1:
+            hvst.subject_trees[subject_tree[0]]["_a"]["_a"]["_a"] ~= doc_harvest;
+            break;
+          case 2:
+            hvst.subject_trees[subject_tree[0]][subject_tree[1]]["_a"]["_a"] ~= doc_harvest;
+            break;
+          case 3:
+            hvst.subject_trees[subject_tree[0]][subject_tree[1]][subject_tree[2]]["_a"] ~= doc_harvest;
+            break;
+          case 4:
+            hvst.subject_trees[subject_tree[0]][subject_tree[1]][subject_tree[2]][subject_tree[3]] ~= doc_harvest;
+            break;
+          default:
+            break;
+          }
+          _topic_register ~= _sub_topic_register.join("\n");
+        }
+        auto char_repeat_number = (doc_harvest.title.length
+          + doc_harvest.author.length + 16);
+        char_repeat_number = (char_repeat_number > min_repeat_number)
+        ? char_repeat_number
+        : min_repeat_number;
+        _document_topic_register ~= format(
+          "\"%s\", %s%s\n%s",
+          doc_harvest.title,
+          doc_harvest.author,
+          (doc_harvest.date_published.length > 0) ? " (" ~ doc_harvest.date_published ~ ")" : "",
+          _topic_register.sort!("toUpper(a) < toUpper(b)", SwapStrategy.unstable).release.join("\n"),
+        );
+      }
+      topics ~= format(q"┃<!DOCTYPE html>
+<html>
+<head>
+<meta charset="utf-8">
+<title>Metadata Harvest - Topics</title>
+<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="doc_reform" />
+<link rel="generator" href="http://sisudoc.org" />
+<link href="../../_sisu/css/harvest.css" rel="stylesheet">
+<style TYPE="text/css">
+/* DocReform harvest css default stylesheet */
+  body {
+    color: black;
+    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;
+  }
+
+  .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;
+  }
+  h1 {
+    font-size: 120%%;
+    font-weight: bold;
+    color: white;
+    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;
+    color: white;
+    background: #880000;
+  }
+  p.lev0 {
+    font-size: 120%%;
+    margin-left: 1em;
+    color: white;
+    background: #000000;
+  }
+
+  p.lev1 {
+    font-size: 110%%;
+    margin-left: 2em;
+    color: white;
+    background: #444444;
+  }
+  p.lev2 {
+    font-size: 100%%;
+    margin-left: 3em;
+    background: #888888;
+  }
+  p.lev3 {
+    font-size: 90%%;
+    margin-left: 4em;
+    background: #bbbbbb;
+  }
+  p.lev4 {
+    font-size: 80%%;
+    margin-left: 5em;
+    background: #eeeeee;
+  }
+  p.lev5 {
+    font-size: 80%%;
+    margin-left: 6em;
+  }
+</style>
+<link rel="shortcut icon" href="../_sisu/image/rb7.ico" />
+</head>
+<body lang="en" xml:lang="en">
+<a name="top" id="top"></a>
+<a name="up" id="up"></a>
+<a name="start" id="start"></a>
+<h1>Metadata Harvest - Topics (output organised by language &amp; filetype)</h1>
+<p>[<a href="../../index.html">&nbsp;HOME&nbsp;</a>] also see <a href="authors.html">Metadata Harvest - Authors</a></p>
+<p><a href="#A">A</a>,&nbsp;<a href="#B">B</a>,&nbsp;<a href="#C">C</a>,&nbsp;<a href="#D">D</a>,&nbsp;<a href="#E">E</a>,&nbsp;<a href="#F">F</a>,&nbsp;<a href="#G">G</a>,&nbsp;<a href="#H">H</a>,&nbsp;<a href="#I">I</a>,&nbsp;<a href="#J">J</a>,&nbsp;<a href="#K">K</a>,&nbsp;<a href="#L">L</a>,&nbsp;<a href="#M">M</a>,&nbsp;<a href="#N">N</a>,&nbsp;<a href="#O">O</a>,&nbsp;<a href="#P">P</a>,&nbsp;<a href="#Q">Q</a>,&nbsp;<a href="#R">R</a>,&nbsp;<a href="#S">S</a>,&nbsp;<a href="#T">T</a>,&nbsp;<a href="#U">U</a>,&nbsp;<a href="#V">V</a>,&nbsp;<a href="#W">W</a>,&nbsp;<a href="#X">X</a>,&nbsp;<a href="#Y">Y</a>,&nbsp;<a href="#Z">Z</a>,&nbsp;
+<p></p>
+<hr />
+<p class="tiny"><a href="../../en/manifest/topics.html">English</a>&nbsp;&nbsp;&nbsp;</p>
+<hr />
+┃") ~ "\n";
+      char _prev_k = "_".to!char;
+      int _kn;
+      foreach(k0;
+        hvst.subject_trees.keys
+        .sort!("toUpper(a) < toUpper(b)", SwapStrategy.unstable)
+      ) {
+        if (k0.toUpper.to!(char[])[0] != _prev_k) {
+          topics ~= format(q"┃<p class="letter"><a name="%s">%s</a></p><p class="book_index_lev1"><a name="a"></a></p>┃",
+            k0.toUpper.to!(char[])[0],
+            k0.toUpper.to!(char[])[0],
+          );
+          _prev_k = k0.toUpper.to!(char[])[0];
+        }
+        if (k0 != "_a") {
+          topics ~= format(q"┃<p class="lev0"><a name="%s">%s</a></p>┃",
+            k0, k0,) ~ "\n";
+          if (_opt_action.very_verbose) {
+            writeln("", k0);
+          }
+          if ("_a" in hvst.subject_trees[k0]) {
+            foreach (t_a_;
+              hvst.subject_trees[k0]["_a"]["_a"]["_a"]
+              .multiSort!("toUpper(a.title) < toUpper(b.title)", "a.author < b.author", SwapStrategy.unstable)
+            ) {
+              _auth = [];
+              if (t_a_.author_arr.length < 2) {
+                _auth = format(q"┃ <a href="authors.html#%s">%s</a>┃",
+                  t_a_.author_surname,
+                  t_a_.author,
+                );
+              } else {
+                foreach (a; t_a_.author_arr) {
+                  _auth ~= format(q"┃ <a href="authors.html#%s">%s</a>,┃",
+                    t_a_.author_surname,
+                    a,
+                  );
+                }
+              }
+              topics ~= format(q"┃<p class="work"><a href="%s">"%s"</a> -%s┃",
+                "url",
+                t_a_.title,
+                _auth,
+              ) ~ "\n";
+              if (_opt_action.very_verbose) {
+                writeln("- ", t_a_.title, " - ", t_a_.author);
+              }
+            }
+          }
+          foreach(k1;
+            hvst.subject_trees[k0].keys
+            .sort!("toUpper(a) < toUpper(b)", SwapStrategy.unstable)
+          ) {
+            if (k1 != "_a") {
+              topics ~= format(q"┃<p class="lev1"><a name="%s">%s</a></p>┃",
+                k1, k1,) ~ "\n";
+              if (_opt_action.very_verbose) {
+                writeln("  ", k1);
+              }
+              if ("_a" in hvst.subject_trees[k0][k1]) {
+                foreach (t_a_;
+                  hvst.subject_trees[k0][k1]["_a"]["_a"]
+                  .multiSort!("toUpper(a.title) < toUpper(b.title)", "a.author < b.author", SwapStrategy.unstable)
+                ) {
+                  _auth = [];
+                  if (t_a_.author_arr.length < 2) {
+                    _auth = format(q"┃ <a href="authors.html#%s">%s</a>┃",
+                      t_a_.author_surname,
+                      t_a_.author,
+                    );
+                  } else {
+                    foreach (a; t_a_.author_arr) {
+                      _auth ~= format(q"┃ <a href="authors.html#%s">%s</a>,┃",
+                        t_a_.author_surname,
+                        a,
+                      );
+                    }
+                  }
+                  topics ~= format(q"┃<p class="work"><a href="%s">%s</a> -%s┃",
+                    "url",
+                    t_a_.title,
+                    _auth,
+                  ) ~ "\n";
+                  if (_opt_action.very_verbose) {
+                    writeln("  - ", t_a_.title, " - ", t_a_.author);
+                  }
+                }
+              }
+            }
+            foreach(k2;
+              hvst.subject_trees[k0][k1].keys
+              .sort!("toUpper(a) < toUpper(b)", SwapStrategy.unstable)
+            ) {
+              if (k2 != "_a") {
+                topics ~= format(q"┃<p class="lev2"><a name="%s">%s</a></p>┃",
+                  k2, k2,) ~ "\n";
+                if (_opt_action.very_verbose) {
+                  writeln("    ", k2);
+                }
+                if ("_a" in hvst.subject_trees[k0][k1][k2]) {
+                  foreach (t_a_;
+                    hvst.subject_trees[k0][k1][k2]["_a"]
+                    .multiSort!("toUpper(a.title) < toUpper(b.title)", "a.author < b.author", SwapStrategy.unstable)
+                  ) {
+                    _auth = [];
+                    if (t_a_.author_arr.length < 2) {
+                      _auth = format(q"┃ <a href="authors.html#%s">%s</a>┃",
+                        t_a_.author_surname,
+                        t_a_.author,
+                      );
+                    } else {
+                      foreach (a; t_a_.author_arr) {
+                        _auth ~= format(q"┃ <a href="authors.html#%s">%s</a>,┃",
+                          t_a_.author_surname,
+                          a,
+                        );
+                      }
+                    }
+                    topics ~= format(q"┃<p class="work"><a href="%s">%s</a> -%s┃",
+                      "url",
+                      t_a_.title,
+                      _auth,
+                    ) ~ "\n";
+                    if (_opt_action.very_verbose) {
+                      writeln("    - ", t_a_.title, " - ", t_a_.author);
+                    }
+                  }
+                }
+              }
+              foreach(k3;
+                hvst.subject_trees[k0][k1][k2].keys
+                .sort!("toUpper(a) < toUpper(b)", SwapStrategy.unstable)
+              ) {
+                if (k3 != "_a") {
+                  topics ~= format(q"┃<p class="lev3"><a name="%s">%s</a></p>┃",
+                    k3, k3,) ~ "\n";
+                  if (_opt_action.very_verbose) {
+                    writeln("      ", k3);
+                  }
+                  {
+                    foreach (t_a_;
+                      hvst.subject_trees[k0][k1][k2][k3]
+                      .multiSort!("toUpper(a.title) < toUpper(b.title)", "a.author < b.author", SwapStrategy.unstable)
+                    ) {
+                      _auth = [];
+                      if (t_a_.author_arr.length < 2) {
+                        _auth = format(q"┃<a href="authors.html#%s">%s</a>┃",
+                          t_a_.author_surname,
+                          t_a_.author,
+                        );
+                      } else {
+                        foreach (a; t_a_.author_arr) {
+                          _auth ~= format(q"┃ <a href="authors.html#%s">%s</a>,┃",
+                            t_a_.author_surname,
+                            a,
+                          );
+                        }
+                      }
+                      topics ~= format(q"┃ <p class="work"><a href="%s">%s</a> -%s┃",
+                        "url",
+                        t_a_.title,
+                        _auth,
+                      ) ~ "\n";
+                      if (_opt_action.very_verbose) {
+                        writeln("      - ", t_a_.title, " - ", t_a_.author);
+                      }
+                    }
+                  }
+                }
+              }
+            }
+          }
+        }
+      }
+      topics
+      ~= 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";
+    try {
+      auto f = File("topics.html", "w");
+      foreach (o; topics) {
+        f.writeln(o);
+      }
+    } catch (ErrnoException ex) {
+      // Handle error
+    }
+  }
+}
author	Ralph Amissah <ralph.amissah@gmail.com>	2019-06-20 13:45:58 -0400
committer	Ralph Amissah <ralph.amissah@gmail.com>	2019-10-17 19:07:20 -0400
commit	2dd7d2083c23fe9e79a984c534c2283fc4d7e581 (patch)
tree	22deaa43f6888ea6c324a27b5fd4ec9b3546fb63 /src/doc_reform/meta/metadoc_harvests_topics.d
parent	harvest html output (diff)