0.7.3 start to look at document harvest (initial stub)

1 files changed, 128 insertions, 87 deletions

diff --git a/src/doc_reform/meta/conf_make_meta_json.d b/src/doc_reform/meta/conf_make_meta_json.d
index fcd52c1..5fd4499 100644
--- a/src/doc_reform/meta/conf_make_meta_json.d
+++ b/src/doc_reform/meta/conf_make_meta_json.d
@@ -5,6 +5,8 @@
 module doc_reform.meta.conf_make_meta_json;
 static template contentJSONtoDocReformStruct() {
   import
+    std.algorithm,
+    std.array,
     std.exception,
     std.regex,
     std.stdio,
@@ -16,6 +18,7 @@ static template contentJSONtoDocReformStruct() {
   import
     doc_reform.meta.conf_make_meta_structs,
     doc_reform.meta.conf_make_meta_json,
+    doc_reform.meta.defaults,
     doc_reform.meta.rgx;
   ConfCompositePlus _struct_composite;
   auto contentJSONtoDocReformStruct(C,J)(C _struct_composite, J _json, string _identifier) {
@@ -361,6 +364,108 @@ static template contentJSONtoDocReformStruct() {
       }
     }
     /+ meta ------------------------------------------------------------------- +/
+    if (_struct_composite.meta.creator_author.empty) {
+      if ("creator" in _json.object) {
+        if ("author" in _json.object["creator"]
+          && (_json.object["creator"]["author"].type().to!string == "string")
+        ) {
+          _struct_composite.meta.creator_author = _json.object["creator"]["author"].str;
+        }
+        if ("email" in _json.object["creator"]
+          && (_json.object["creator"]["email"].type().to!string == "string")
+        ) {
+          _struct_composite.meta.creator_author_email = _json.object["creator"]["email"].str;
+        }
+        if ("illustrator" in _json.object["creator"]
+          && (_json.object["creator"]["illustrator"].type().to!string == "string")
+        ) {
+          _struct_composite.meta.creator_illustrator = _json.object["creator"]["illustrator"].str;
+        }
+        if ("translator" in _json.object["creator"]
+          && (_json.object["creator"]["translator"].type().to!string == "string")
+        ) {
+          _struct_composite.meta.creator_translator = _json.object["creator"]["translator"].str;
+        }
+      }
+      string[] authors_arr;
+      string[][string] authors_hash_arr = [ "first" : [], "last" : [], "full" : [], "last_first" : [], "as_input" : [] ];
+      string[] authors_raw_arr
+        = _struct_composite.meta.creator_author.split(rgx.arr_delimiter);
+      auto _lastname = appender!(char[])();
+      foreach (author_raw; authors_raw_arr) {
+        authors_arr                  ~= author_raw.replace(rgx.raw_author_munge, "$2 $1");
+        authors_hash_arr["first"]    ~= author_raw.replace(rgx.raw_author_munge, "$2");
+        authors_hash_arr["last"]     ~= author_raw.replace(rgx.raw_author_munge, "$1");
+        authors_hash_arr["full"]     ~= author_raw.replace(rgx.raw_author_munge, "$2 $1");
+        authors_hash_arr["as_input"] ~= author_raw;
+        if (auto m = author_raw.match(rgx.raw_author_munge)) {
+          (m.captures[1]).map!toUpper.copy(_lastname);
+          authors_hash_arr["last_first"] ~= _lastname.data.to!string ~ ", " ~ m.captures[2];
+          _lastname = appender!(char[])();
+        }
+      }
+      _struct_composite.meta.creator_author     = authors_arr.join(", ").chomp.chomp;
+      string _author_name_last_first = authors_hash_arr["last_first"].join("; ").chomp.chomp;
+      _struct_composite.meta.creator_author_surname_fn = (_author_name_last_first.length > 0)
+      ? _author_name_last_first
+      : authors_hash_arr["as_input"].join("; ").chomp.chomp;
+    }
+    if (_struct_composite.meta.title_main.empty) {
+      if ("title" in _json.object) {
+        if ((_json.object["title"].type().to!string) == "string") {
+          _struct_composite.meta.title_main = _json.object["title"].str;
+        } else {
+          if ("edition" in _json.object["title"]
+            && (_json.object["title"]["edition"].type().to!string == "string")
+          ) {
+            _struct_composite.meta.title_edition = _json.object["title"]["edition"].str;
+          }
+          if ("full" in _json.object["title"]
+            && (_json.object["title"]["full"].type().to!string == "string")
+          ) {}
+          if ("language" in _json.object["title"]
+            && (_json.object["title"]["language"].type().to!string == "string")
+          ) {
+            _struct_composite.meta.title_language = _json.object["title"]["language"].str;
+          }
+          if ("main" in _json.object["title"]
+            && (_json.object["title"]["main"].type().to!string == "string")
+          ) {
+            _struct_composite.meta.title_main = _json.object["title"]["main"].str;
+          } else if ("title" in _json.object["title"]
+            && (_json.object["title"]["title"].type().to!string == "string")
+          ) {
+            _struct_composite.meta.title_main = _json.object["title"]["title"].str;
+          }
+          if ("note" in _json.object["title"]
+            && (_json.object["title"]["note"].type().to!string == "string")
+          ) {
+            _struct_composite.meta.title_note = _json.object["title"]["note"].str;
+          }
+          if ("sub" in _json.object["title"]
+            && (_json.object["title"]["sub"].type().to!string == "string")
+          ) {
+            _struct_composite.meta.title_sub = _json.object["title"]["sub"].str;
+          }
+          if ("subtitle" in _json.object["title"]
+            && (_json.object["title"]["subtitle"].type().to!string == "string")
+          ) {
+            _struct_composite.meta.title_subtitle = _json.object["title"]["subtitle"].str;
+          }
+        }
+      }
+      if ((!(_struct_composite.meta.title_subtitle.empty))
+      && (_struct_composite.meta.title_sub.empty)) {
+        _struct_composite.meta.title_sub = _struct_composite.meta.title_subtitle;
+      }
+      _struct_composite.meta.title_full = (_struct_composite.meta.title_sub.empty)
+      ? _struct_composite.meta.title_main
+      : format(
+          "%s - %s",
+          _struct_composite.meta.title_main,
+          _struct_composite.meta.title_sub,
+        );
+    }
     if ("classify" in _json.object) {
       if ("dewey" in _json.object["classify"]
         && (_json.object["classify"]["dewey"].type().to!string == "string")
@@ -386,6 +491,22 @@ static template contentJSONtoDocReformStruct() {
         && (_json.object["classify"]["topic_register"].type().to!string == "string")
       ) {
         _struct_composite.meta.classify_topic_register = _json.object["classify"]["topic_register"].str;
+        string[] main_topics_ = _struct_composite.meta.classify_topic_register.split(rgx.topic_register_main_terms_split);
+        string[] topics;
+        string   topics_tmp;
+        string[] multiple_sub_terms;
+        foreach (mt; main_topics_) {
+          topics_tmp = mt.replaceAll(rgx.topic_register_main_term_plus_rest_split,    mkup.sep);
+          if (auto m = topics_tmp.match(rgx.topic_register_multiple_sub_terms_split)) {
+            multiple_sub_terms = m.captures[1].split(rgx.topic_register_sub_terms_split);
+            foreach (subterm; multiple_sub_terms) {
+              topics ~= m.captures.pre ~ mkup.sep ~ subterm;
+            }
+          } else {
+            topics ~= topics_tmp;
+          }
+        }
+        _struct_composite.meta.classify_topic_register_arr = topics;
       }
     }
     if ("date" in _json.object) {
@@ -424,6 +545,13 @@ static template contentJSONtoDocReformStruct() {
       ) {
         _struct_composite.meta.date_valid = _json.object["date"]["valid"].str;
       }
+      _struct_composite.meta.author_date_title = format(
+        "%s %s \"%s\"",
+        _struct_composite.meta.creator_author_surname_fn,
+        (_struct_composite.meta.date_published.length > 0)
+          ? "(" ~ _struct_composite.meta.date_published ~ ")" : "",
+        _struct_composite.meta.title_full,
+      );
     }
     if ("links" in _json.object) {}
     if ("notes" in _json.object) {
@@ -508,93 +636,6 @@ static template contentJSONtoDocReformStruct() {
         _struct_composite.meta.rights_license = _json.object["rights"]["license"].str;
       }
     }
-    if (_struct_composite.meta.creator_author.empty) {
-      if ("creator" in _json.object) {
-        if ("author" in _json.object["creator"]
-          && (_json.object["creator"]["author"].type().to!string == "string")
-        ) {
-          _struct_composite.meta.creator_author = _json.object["creator"]["author"].str;
-        }
-        if ("email" in _json.object["creator"]
-          && (_json.object["creator"]["email"].type().to!string == "string")
-        ) {
-          _struct_composite.meta.creator_author_email = _json.object["creator"]["email"].str;
-        }
-        if ("illustrator" in _json.object["creator"]
-          && (_json.object["creator"]["illustrator"].type().to!string == "string")
-        ) {
-          _struct_composite.meta.creator_illustrator = _json.object["creator"]["illustrator"].str;
-        }
-        if ("translator" in _json.object["creator"]
-          && (_json.object["creator"]["translator"].type().to!string == "string")
-        ) {
-          _struct_composite.meta.creator_translator = _json.object["creator"]["translator"].str;
-        }
-      }
-      string[] authors_arr;
-      string[] authors_raw_arr
-        = _struct_composite.meta.creator_author.split(rgx.arr_delimiter);
-      foreach (author_raw; authors_raw_arr) {
-        authors_arr ~= author_raw.replace(rgx.raw_author_munge, "$2 $1");
-      }
-      _struct_composite.meta.creator_author = join(authors_arr, ", ").chomp.chomp;
-    }
-    if (_struct_composite.meta.title_main.empty) {
-      if ("title" in _json.object) {
-        if ((_json.object["title"].type().to!string) == "string") {
-          _struct_composite.meta.title_main = _json.object["title"].str;
-        } else {
-          if ("edition" in _json.object["title"]
-            && (_json.object["title"]["edition"].type().to!string == "string")
-          ) {
-            _struct_composite.meta.title_edition = _json.object["title"]["edition"].str;
-          }
-          if ("full" in _json.object["title"]
-            && (_json.object["title"]["full"].type().to!string == "string")
-          ) {}
-          if ("language" in _json.object["title"]
-            && (_json.object["title"]["language"].type().to!string == "string")
-          ) {
-            _struct_composite.meta.title_language = _json.object["title"]["language"].str;
-          }
-          if ("main" in _json.object["title"]
-            && (_json.object["title"]["main"].type().to!string == "string")
-          ) {
-            _struct_composite.meta.title_main = _json.object["title"]["main"].str;
-          } else if ("title" in _json.object["title"]
-            && (_json.object["title"]["title"].type().to!string == "string")
-          ) {
-            _struct_composite.meta.title_main = _json.object["title"]["title"].str;
-          }
-          if ("note" in _json.object["title"]
-            && (_json.object["title"]["note"].type().to!string == "string")
-          ) {
-            _struct_composite.meta.title_note = _json.object["title"]["note"].str;
-          }
-          if ("sub" in _json.object["title"]
-            && (_json.object["title"]["sub"].type().to!string == "string")
-          ) {
-            _struct_composite.meta.title_sub = _json.object["title"]["sub"].str;
-          }
-          if ("subtitle" in _json.object["title"]
-            && (_json.object["title"]["subtitle"].type().to!string == "string")
-          ) {
-            _struct_composite.meta.title_subtitle = _json.object["title"]["subtitle"].str;
-          }
-        }
-      }
-      if ((!(_struct_composite.meta.title_subtitle.empty))
-      && (_struct_composite.meta.title_sub.empty)) {
-        _struct_composite.meta.title_sub = _struct_composite.meta.title_subtitle;
-      }
-      _struct_composite.meta.title_full = (_struct_composite.meta.title_sub.empty)
-      ? _struct_composite.meta.title_main
-      : format(
-          "%s - %s",
-          _struct_composite.meta.title_main,
-          _struct_composite.meta.title_sub,
-        );
-    }
     return _struct_composite;
   }
 }