Merge pull request apache#52 from michellethomas/cherry_pick_annotati…

…ons_empty_metrics

Cherry pick annotations empty metrics

superset/assets/.istanbul.yml

@@ -1,6 +1,6 @@
 verbose: true
 instrumentation:
-    root: './javascripts'
+    root: './src'
     extensions: ['.js', '.jsx']
     excludes: [
         'dist/**'

superset/assets/src/explore/controls.jsx

@@ -959,12 +959,14 @@ export const controls = {
   },
 
   timeseries_limit_metric: {
-    type: 'SelectControl',
+    type: 'MetricsControl',
     label: t('Sort By'),
     default: null,
     description: t('Metric used to define the top series'),
     mapStateToProps: state => ({
-      choices: (state.datasource) ? state.datasource.metrics_combo : [],
+      columns: state.datasource ? state.datasource.columns : [],
+      savedMetrics: state.datasource ? state.datasource.metrics : [],
+      datasourceType: state.datasource && state.datasource.type,
     }),
   },
 

superset/assets/src/visualizations/heatmap.js

@@ -177,11 +177,12 @@ function heatmapVis(slice, payload) {
       const k = d3.mouse(this);
       const m = Math.floor(scale[0].invert(k[0]));
       const n = Math.floor(scale[1].invert(k[1]));
+      const metric = typeof fd.metric === 'object' ? fd.metric.label : fd.metric;
       if (m in matrix && n in matrix[m]) {
         const obj = matrix[m][n];
         s += '<div><b>' + fd.all_columns_x + ': </b>' + obj.x + '<div>';
         s += '<div><b>' + fd.all_columns_y + ': </b>' + obj.y + '<div>';
-        s += '<div><b>' + fd.metric + ': </b>' + valueFormatter(obj.v) + '<div>';
+        s += '<div><b>' + metric + ': </b>' + valueFormatter(obj.v) + '<div>';
         if (fd.show_perc) {
           s += '<div><b>%: </b>' + fp(fd.normalized ? obj.rank : obj.perc) + '<div>';
         }

superset/assets/src/visualizations/nvd3_vis.js

@@ -784,6 +784,12 @@ export default function nvd3Vis(slice, payload) {
             });
           });
         }
+
+        // rerender chart appended with annotation layer
+        svg.datum(data)
+          .attr('height', height)
+          .attr('width', width)
+          .call(chart);
       }
     }
     return chart;

superset/assets/src/visualizations/table.js

@@ -17,7 +17,7 @@ function tableVis(slice, payload) {
   const data = payload.data;
   const fd = slice.formData;
 
-  let metrics = fd.metrics || [];
+  let metrics = fd.metrics.map(m => m.label || m);
   // Add percent metrics
   metrics = metrics.concat((fd.percent_metrics || []).map(m => '%' + m));
   // Removing metrics (aggregates) that are strings
@@ -187,15 +187,15 @@ function tableVis(slice, payload) {
   let sortBy;
   if (fd.timeseries_limit_metric) {
     // Sort by as specified
-    sortBy = fd.timeseries_limit_metric;
+    sortBy = fd.timeseries_limit_metric.label || fd.timeseries_limit_metric;
   } else if (metrics.length > 0) {
     // If not specified, use the first metric from the list
     sortBy = metrics[0];
   }
   if (sortBy) {
     datatable.column(data.columns.indexOf(sortBy)).order(fd.order_desc ? 'desc' : 'asc');
   }
-  if (fd.timeseries_limit_metric && metrics.indexOf(fd.timeseries_limit_metric) < 0) {
+  if (sortBy && metrics.indexOf(sortBy) < 0) {
     // Hiding the sortBy column if not in the metrics list
     datatable.column(data.columns.indexOf(sortBy)).visible(false);
   }

superset/connectors/druid/models.py

@@ -1098,6 +1098,18 @@ def _dimensions_to_values(dimensions):
 
         return values
 
+    @staticmethod
+    def sanitize_metric_object(metric):
+        """
+        Update a metric with the correct type if necessary.
+        :param dict metric: The metric to sanitize
+        """
+        if (
+            utils.is_adhoc_metric(metric) and
+            metric['column']['type'].upper() == 'FLOAT'
+        ):
+            metric['column']['type'] = 'DOUBLE'
+
     def run_query(  # noqa / druid
             self,
             groupby, metrics,
@@ -1141,11 +1153,8 @@ def run_query(  # noqa / druid
             LooseVersion(self.cluster.get_druid_version()) < LooseVersion('0.11.0')
         ):
             for metric in metrics:
-                if (
-                    utils.is_adhoc_metric(metric) and
-                    metric['column']['type'].upper() == 'FLOAT'
-                ):
-                    metric['column']['type'] = 'DOUBLE'
+                self.sanitize_metric_object(metric)
+            self.sanitize_metric_object(timeseries_limit_metric)
 
         aggregations, post_aggs = DruidDatasource.metrics_and_post_aggs(
             metrics,
@@ -1201,7 +1210,7 @@ def run_query(  # noqa / druid
             logging.info('Running two-phase topn query for dimension [{}]'.format(dim))
             pre_qry = deepcopy(qry)
             if timeseries_limit_metric:
-                order_by = timeseries_limit_metric
+                order_by = utils.get_metric_name(timeseries_limit_metric)
                 aggs_dict, post_aggs_dict = DruidDatasource.metrics_and_post_aggs(
                     [timeseries_limit_metric],
                     metrics_dict)
@@ -1270,7 +1279,7 @@ def run_query(  # noqa / druid
                     order_by = pre_qry_dims[0]
 
                 if timeseries_limit_metric:
-                    order_by = timeseries_limit_metric
+                    order_by = utils.get_metric_name(timeseries_limit_metric)
                     aggs_dict, post_aggs_dict = DruidDatasource.metrics_and_post_aggs(
                         [timeseries_limit_metric],
                         metrics_dict)

superset/connectors/sqla/models.py

@@ -651,6 +651,8 @@ def get_sqla_query(  # sqla
 
         for col, ascending in orderby:
             direction = asc if ascending else desc
+            if utils.is_adhoc_metric(col):
+                col = self.adhoc_metric_to_sa(col, cols)
             qry = qry.order_by(direction(col))
 
         if row_limit:
@@ -675,8 +677,15 @@ def get_sqla_query(  # sqla
 
                 ob = inner_main_metric_expr
                 if timeseries_limit_metric:
-                    timeseries_limit_metric = metrics_dict.get(timeseries_limit_metric)
-                    ob = timeseries_limit_metric.sqla_col
+                    if utils.is_adhoc_metric(timeseries_limit_metric):
+                        ob = self.adhoc_metric_to_sa(timeseries_limit_metric, cols)
+                    elif timeseries_limit_metric in metrics_dict:
+                        timeseries_limit_metric = metrics_dict.get(
+                            timeseries_limit_metric,
+                        )
+                        ob = timeseries_limit_metric.sqla_col
+                    else:
+                        raise Exception(_("Metric '{}' is not valid".format(m)))
                 direction = desc if order_desc else asc
                 subq = subq.order_by(direction(ob))
                 subq = subq.limit(timeseries_limit)

superset/data/__init__.py

@@ -626,7 +626,8 @@ def load_birth_names():
                     'op': 'in',
                     'val': ['girl'],
                 }],
-                row_limit=50)),
+                row_limit=50,
+                timeseries_limit_metric='sum__num')),
         Slice(
             slice_name="Boys",
             viz_type='table',
@@ -760,6 +761,53 @@ def load_birth_names():
                 },
                 viz_type="big_number_total",
                 granularity_sqla="ds")),
+        Slice(
+            slice_name='Top 10 California Names Timeseries',
+            viz_type='line',
+            datasource_type='table',
+            datasource_id=tbl.id,
+            params=get_slice_json(
+                defaults,
+                metrics=[{
+                    'expressionType': 'SIMPLE',
+                    'column': {
+                        'column_name': 'num_california',
+                        'expression': "CASE WHEN state = 'CA' THEN num ELSE 0 END",
+                    },
+                    'aggregate': 'SUM',
+                    'label': 'SUM(num_california)',
+                }],
+                viz_type='line',
+                granularity_sqla='ds',
+                groupby=['name'],
+                timeseries_limit_metric={
+                    'expressionType': 'SIMPLE',
+                    'column': {
+                        'column_name': 'num_california',
+                        'expression': "CASE WHEN state = 'CA' THEN num ELSE 0 END",
+                    },
+                    'aggregate': 'SUM',
+                    'label': 'SUM(num_california)',
+                },
+                limit='10')),
+        Slice(
+            slice_name="Names Sorted by Num in California",
+            viz_type='table',
+            datasource_type='table',
+            datasource_id=tbl.id,
+            params=get_slice_json(
+                defaults,
+                groupby=['name'],
+                row_limit=50,
+                timeseries_limit_metric={
+                    'expressionType': 'SIMPLE',
+                    'column': {
+                        'column_name': 'num_california',
+                        'expression': "CASE WHEN state = 'CA' THEN num ELSE 0 END",
+                    },
+                    'aggregate': 'SUM',
+                    'label': 'SUM(num_california)',
+                })),
     ]
     for slc in slices:
         merge_slice(slc)

superset/viz.py

@@ -59,6 +59,7 @@ class BaseViz(object):
     is_timeseries = False
     default_fillna = 0
     cache_type = 'df'
+    enforce_numerical_metrics = True
 
     def __init__(self, datasource, form_data, force=False):
         if not datasource:
@@ -205,7 +206,8 @@ def get_df(self, query_obj=None):
                     df[DTTM_ALIAS] += timedelta(hours=self.datasource.offset)
                 df[DTTM_ALIAS] += self.time_shift
 
-            self.df_metrics_to_num(df, query_obj.get('metrics') or [])
+            if self.enforce_numerical_metrics:
+                self.df_metrics_to_num(df, query_obj.get('metrics') or [])
 
             df.replace([np.inf, -np.inf], np.nan)
             fillna = self.get_fillna_for_columns(df.columns)
@@ -480,6 +482,7 @@ class TableViz(BaseViz):
     verbose_name = _('Table View')
     credits = 'a <a href="https://github.com/airbnb/superset">Superset</a> original'
     is_timeseries = False
+    enforce_numerical_metrics = False
 
     def should_be_timeseries(self):
         fd = self.form_data
@@ -510,7 +513,8 @@ def query_obj(self):
             order_by_cols = fd.get('order_by_cols') or []
             d['orderby'] = [json.loads(t) for t in order_by_cols]
         elif sort_by:
-            if sort_by not in d['metrics']:
+            sort_by_label = utils.get_metric_name(sort_by)
+            if sort_by_label not in utils.get_metric_names(d['metrics']):
                 d['metrics'] += [sort_by]
             d['orderby'] = [(sort_by, not fd.get('order_desc', True))]
 
@@ -552,7 +556,7 @@ def get_data(self, df):
                 m_name = '%' + m
                 df[m_name] = pd.Series(metric_percents[m], name=m_name)
             # Remove metrics that are not in the main metrics list
-            metrics = fd.get('metrics', [])
+            metrics = fd.get('metrics') or []
             metrics = [self.get_metric_label(m) for m in metrics]
             for m in filter(
                 lambda m: m not in metrics and m in df.columns,