From 7dae35e722d672bfc93892403fecee8dc57d5196 Mon Sep 17 00:00:00 2001
From: Olivier Mangin <olivier@mangin.com>
Date: Thu, 9 Aug 2018 20:54:44 +0200
Subject: [PATCH] Adds the statistics command. (Fixes #8)

---
 pubs/commands/__init__.py       |  6 +++---
 pubs/commands/statistics_cmd.py | 33 +++++++++++++++++++++++++++++++++
 pubs/pubs_cmd.py                |  5 +++--
 tests/test_usecase.py           | 15 +++++++++++++++
 4 files changed, 54 insertions(+), 5 deletions(-)
 create mode 100644 pubs/commands/statistics_cmd.py

diff --git a/pubs/commands/__init__.py b/pubs/commands/__init__.py
index cd17fa3..ed5defa 100644
--- a/pubs/commands/__init__.py
+++ b/pubs/commands/__init__.py
@@ -6,9 +6,11 @@ from . import add_cmd
 from . import rename_cmd
 from . import remove_cmd
 from . import list_cmd
+from . import edit_cmd
+from . import tag_cmd
+from . import statistics_cmd
 # doc
 from . import doc_cmd
-from . import tag_cmd
 from . import note_cmd
 # bulk
 from . import export_cmd
@@ -16,5 +18,3 @@ from . import import_cmd
 # bonus
 from . import websearch_cmd
 from . import url_cmd
-
-from . import edit_cmd
diff --git a/pubs/commands/statistics_cmd.py b/pubs/commands/statistics_cmd.py
new file mode 100644
index 0000000..5132010
--- /dev/null
+++ b/pubs/commands/statistics_cmd.py
@@ -0,0 +1,33 @@
+from ..repo import Repository
+from ..uis import get_ui
+from .. import color
+
+
+def parser(subparsers, conf):
+    parser = subparsers.add_parser(
+        'statistics',
+        help="show statistics on the repository.")
+    return parser
+
+
+def command(conf, args):
+    ui = get_ui()
+    rp = Repository(conf)
+    papers = list(rp.all_papers())
+
+    paper_count = len(papers)
+    doc_count = sum([0 if p.docpath is None else 1 for p in papers])
+    tag_count = len(list(rp.get_tags()))
+    papers_with_tags = sum([0 if p.tags else 1 for p in papers])
+
+    ui.message(color.dye_out('Repository statistics:', 'bold'))
+    ui.message('Total papers: {}, {} ({}) have a document attached'.format(
+        color.dye_out('{:d}'.format(paper_count), 'bgreen'),
+        color.dye_out('{:d}'.format(doc_count), 'bold'),
+        '{:.0f}%'.format(100. * doc_count / paper_count),
+    ))
+    ui.message('Total tags: {}, {} ({}) of papers have at least one tag'.format(
+        color.dye_out('{:d}'.format(tag_count), 'bgreen'),
+        color.dye_out('{:d}'.format(papers_with_tags), 'bold'),
+        '{:.0f}%'.format(100. * papers_with_tags / paper_count),
+    ))
diff --git a/pubs/pubs_cmd.py b/pubs/pubs_cmd.py
index 949fd4a..258cb85 100644
--- a/pubs/pubs_cmd.py
+++ b/pubs/pubs_cmd.py
@@ -20,9 +20,11 @@ CORE_CMDS = collections.OrderedDict([
     ('rename', commands.rename_cmd),
     ('remove', commands.remove_cmd),
     ('list', commands.list_cmd),
+    ('edit', commands.edit_cmd),
+    ('tag', commands.tag_cmd),
+    ('statistics', commands.statistics_cmd),
 
     ('doc', commands.doc_cmd),
-    ('tag', commands.tag_cmd),
     ('note', commands.note_cmd),
 
     ('export', commands.export_cmd),
@@ -30,7 +32,6 @@ CORE_CMDS = collections.OrderedDict([
 
     ('websearch', commands.websearch_cmd),
     ('url', commands.url_cmd),
-    ('edit', commands.edit_cmd),
 ])
 
 
diff --git a/tests/test_usecase.py b/tests/test_usecase.py
index 43cbf2c..11cce7a 100644
--- a/tests/test_usecase.py
+++ b/tests/test_usecase.py
@@ -925,6 +925,21 @@ class TestUsecase(DataCommandTestCase):
         self.assertFalse(os.path.isfile(self.default_conf_path))
         self.assertTrue(os.path.isfile(alt_conf))
 
+    def test_statistics(self):
+        cmds = ['pubs init',
+                'pubs add data/pagerank.bib',
+                'pubs add -d data/turing-mind-1950.pdf data/turing1950.bib',
+                'pubs add data/martius.bib',
+                'pubs add data/10.1371%2Fjournal.pone.0038236.bib',
+                'pubs tag Page99 A+B',
+                'pubs tag turing1950computing C',
+                'pubs statistics',
+                ]
+        out = self.execute_cmds(cmds)
+        lines = out[-1].splitlines()
+        self.assertEqual(lines[0], 'Repository statistics:')
+        self.assertEqual(lines[1], 'Total papers: 4, 1 (25%) have a document attached')
+        self.assertEqual(lines[2], 'Total tags: 3, 2 (50%) of papers have at least one tag')
 
 
 @ddt.ddt