Automagically write yaml config file for each individual dataset.

diff --git a/gerritstats/classes.py b/gerritstats/classes.py
index fe8835b..7807c88 100644 (file)
--- a/gerritstats/classes.py
+++ b/gerritstats/classes.py
@@ -21,6 +21,7 @@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
 
 import os
 from datetime import datetime
+from cStringIO import StringIO
 
 class Metric(object):
     '''
@@ -67,6 +68,9 @@ class Settings(object):
     
     def __str__(self):
         return 'Metrics container object'
+    
+    def get_measures(self):
+        return self.queries.keys()
 
 
 class Gerrit(object):
@@ -110,10 +114,21 @@ class Repo(object):
             self.dataset[metric]['volunteer'] = 0
             self.dataset[metric]['total'] = 0
             self.num_metrics +=1
+        
+        self.labels = self.set_labels(settings)
     
     def __str__(self):
         return self.name
     
+    def set_labels(self, settings):
+        labels = []
+        for metric in settings.metrics:
+            headings = self.dataset[metric]
+            headings = ['%s_%s' % (metric, heading) for heading in headings]
+            labels.extend(headings)
+        return labels
+            
+
     def determine_directory(self):
         return os.path.join(self.gerrit.data_location, self.name)
 
@@ -134,3 +149,67 @@ class Repo(object):
         else:
             return 'a'
 
+
+class YamlConfig(object):
+    def __init__(self, settings, repo):
+        self.format = 'csv'
+        self.repo = repo
+        self.dataset_id = repo.name
+        self.name = self.set_description(settings, repo)
+        self.shortname = 'Codereview stats for %s' % repo.name
+        self.url = self.set_url()
+        self.buffer = StringIO()
+    
+    def set_url(self):
+        return '/data/datasources/gerrit-stats/%s.%s' % (self.dataset_id, self.format)
+    
+    def set_description(self, settings, repo):
+        measures = ','.join(settings.get_measures())
+        return '%s metrics for %s repo' % (measures, repo.name)
+    
+    def set_metadata(self):
+        self.buffer.write('id: %s\n' % self.dataset_id)
+        self.buffer.write('name: %s\n' % self.name)
+        self.buffer.write('shortName: %s\n' % self.shortname)
+        self.buffer.write('format: %s\n' % self.format)
+        self.buffer.write('url: %s\n' % self.url)
+        self.buffer.write('\n')
+    
+    def set_timespan(self):
+        today = datetime.today()
+        self.buffer.write('timespan:\n')
+        self.buffer.write('    end: %s/%s/%s\n' % (today.year, today.month, today.day))
+        self.buffer.write('    start: 2012/6/12\n')
+        self.buffer.write('    step: 1d\n')
+        self.buffer.write('\n')
+        
+    def set_columns(self, labels):
+        num_labels = len(labels)
+        self.buffer.write('columns:\n')
+        self.buffer.write('    labels:\n')
+        self.buffer.write('    - Date\n')
+        for label in labels:
+            self.buffer.write('    - %s\n' % label)
+        self.buffer.write('    types:\n')
+        self.buffer.write('    - date\n')
+        for x in xrange(num_labels):
+            self.buffer.write('    - int\n')
+        self.buffer.write('\n')
+    
+    def set_charttype(self):
+        self.buffer.write('chart:\n')
+        self.buffer.write('    chartType: dygraphs\n')
+    
+    def write_file(self):
+        self.set_metadata()
+        self.set_timespan()
+        self.set_columns(self.repo.labels)
+        self.set_charttype()
+        
+        filename = '%s.yaml' % (self.repo.determine_filename())
+        print self.repo.directory
+        full_path = os.path.join(self.repo.directory, filename)
+        print 'filename: %s' % filename
+        fh = open(full_path, 'w')
+        fh.write(self.buffer.getvalue())
+        fh.close()
\ No newline at end of file

diff --git a/gerritstats/stats.py b/gerritstats/stats.py
index 2517ad1..0866e5d 100755 (executable)
--- a/gerritstats/stats.py
+++ b/gerritstats/stats.py
@@ -24,7 +24,7 @@ import sys
 import os
 from datetime import datetime
 
-from classes import Gerrit, Settings, Metric, Repo
+from classes import Gerrit, Settings, Metric, Repo, YamlConfig
 
 def create_repo_set(gerrit, settings):
     repos = {}
@@ -62,6 +62,7 @@ def output_results(fh, *args):
     fh.write(output)
     sys.stdout.write(output)
 
+
 def write_heading(fh, repo):
     output_results(fh, 'data',',','repository',',')
     for metric_counter, (name, metric) in enumerate(repo.dataset.iteritems()):
@@ -86,8 +87,10 @@ def run_gerrit_query(query):
     return output
 
 
-def create_dataset(repos, gerrit):
+def create_dataset(repos, gerrit, settings):
     for key, repo in repos.iteritems():
+        yaml = YamlConfig(settings, repo)
+        yaml.write_file()
         fh = open(repo.full_path, repo.filemode)       
         if repo.filemode == 'w':
             write_heading(fh, repo)
@@ -157,6 +160,7 @@ def construct_dataset(settings, repos, metric, output, gerrit):
                 repo.dataset[metric]['volunteer'] +=1
                 repo.email['volunteer'].add(email)
             repo.touched = True
+            
 
 
 def main():
@@ -166,15 +170,14 @@ def main():
     repos = create_repo_set(gerrit, settings)
     
     for metric in settings.metrics.itervalues():
-        #query = 'ssh -p %s %s gerrit query --format=%s %s' % (gerrit.port, gerrit.host, gerrit.format, question)
-        output = run_gerrit_query(metric.query)
         print 'Running %s' % metric.query
+        output = run_gerrit_query(metric.query)
         construct_dataset(settings, repos, metric.name, output, gerrit)
     
     print 'Fixing miscategorization of volunteer engineers...'
     repos = cleanup_volunteers(repos, settings.whitelist)
     print 'Creating datasets...'
-    create_dataset(repos, gerrit)
+    create_dataset(repos, gerrit, settings)
 
 
 if __name__== '__main__':