chocula: crude publisher type bucketing; field cleanup

1 files changed, 30 insertions, 20 deletions

diff --git a/chocula_schema.sql b/chocula_schema.sql
index 25a2ae3..e7e857a 100644
--- a/chocula_schema.sql
+++ b/chocula_schema.sql
@@ -9,17 +9,26 @@ CREATE TABLE IF NOT EXISTS journal
      publisher TEXT,
      country TEXT,
      lang TEXT,
+
+     release_count INTEGER,
+     ia_count INTEGER,
+     ia_frac FLOAT,
+     kbart_count INTEGER,
+     kbart_frac FLOAT,
+     preserved_count INTEGER,
+     preserved_frac FLOAT,
+
      publisher_type TEXT,
      is_active BOOLEAN,
      is_oa BOOLEAN default false,
      is_longtail BOOLEAN default false,
-     vor_pdf BOOLEAN,
-     vor_html BOOLEAN,
-     vor_jats BOOLEAN,
+     --vor_pdf BOOLEAN,
+     --vor_html BOOLEAN,
+     --vor_jats BOOLEAN,
      has_dois BOOLEAN,
+     any_homepage BOOLEAN,
      any_live_homepage BOOLEAN,
-     any_preservation BOOLEAN,
-     any_ia BOOLEAN
+     bad_issnl BOOLEAN
     );
 
 CREATE TABLE IF NOT EXISTS directory
@@ -27,7 +36,6 @@ CREATE TABLE IF NOT EXISTS directory
      slug TEXT NOT NULL,
      identifier TEXT,
      name TEXT,
-     work_count INTEGER,
      extra TEXT,
      PRIMARY KEY(issnl, slug)
     );
@@ -68,21 +76,23 @@ CREATE TABLE IF NOT EXISTS homepage
      scope TEXT,
      has_issn BOOLEAN,
      blocked BOOLEAN,
+     latest_gwb_success TEXT,
      UNIQUE(issnl, surt)
     );
 CREATE INDEX IF NOT EXISTS homepage_url_idx ON homepage(url);
-    
-CREATE TABLE IF NOT EXISTS fulltext_pattern
-    (id INTEGER PRIMARY KEY,
-     issnl TEXT NOT NULL,
-     surt_glob TEXT,
-     mimetype TEXT,
-     state_date TEXT,
-     end_date TEXT
-    );
 
-CREATE TABLE IF NOT EXISTS title 
-    (id INTEGER PRIMARY KEY,
-     issnl TEXT NOT NULL,
-     title TEXT NOT NULL
-    );
+-- CREATE TABLE IF NOT EXISTS name
+--     (id INTEGER PRIMARY KEY,
+--      issnl TEXT NOT NULL,
+--      name TEXT NOT NULL
+--      -- sluggy (lower-case, no special chars)
+--     );
+    
+-- CREATE TABLE IF NOT EXISTS fulltext_pattern
+--     (id INTEGER PRIMARY KEY,
+--      issnl TEXT NOT NULL,
+--      surt_glob TEXT,
+--      mimetype TEXT,
+--      state_date TEXT,
+--      end_date TEXT
+--     );