feat(db_engine_specs): big query cost estimation (#21325)

Co-authored-by: zamar roura <zamar.roura@cabify.es>
Co-authored-by: Zamar Roura <zamarfazal@gmail.com>

superset-frontend/src/SqlLab/reducers/sqlLab.js

@@ -320,7 +320,7 @@ export default function sqlLabReducer(state = {}, action) {
         ...state,
         queryCostEstimates: {
           ...state.queryCostEstimates,
-          [action.query.sqlEditorId]: {
+          [action.query.id]: {
             completed: false,
             cost: null,
             error: null,
@@ -333,7 +333,7 @@ export default function sqlLabReducer(state = {}, action) {
         ...state,
         queryCostEstimates: {
           ...state.queryCostEstimates,
-          [action.query.sqlEditorId]: {
+          [action.query.id]: {
             completed: true,
             cost: action.json,
             error: null,
@@ -346,7 +346,7 @@ export default function sqlLabReducer(state = {}, action) {
         ...state,
         queryCostEstimates: {
           ...state.queryCostEstimates,
-          [action.query.sqlEditorId]: {
+          [action.query.id]: {
             completed: false,
             cost: null,
             error: action.error,

superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx

@@ -171,7 +171,7 @@ const ExtraOptions = ({
                 />
                 <InfoTooltip
                   tooltip={t(
-                    'For Presto and Postgres, shows a button to compute cost before running a query.',
+                    'For Bigquery, Presto and Postgres, shows a button to compute cost before running a query.',
                   )}
                 />
               </div>

superset/config.py

@@ -478,6 +478,11 @@ def _try_json_readsha(filepath: str, length: int) -> Optional[str]:
     "DRILL_TO_DETAIL": False,
     "DATAPANEL_CLOSED_BY_DEFAULT": False,
     "HORIZONTAL_FILTER_BAR": False,
+    # The feature is off by default, and currently only supported in Presto and Postgres,
+    # and Bigquery.
+    # It also needs to be enabled on a per-database basis, by adding the key/value pair
+    # `cost_estimate_enabled: true` to the database `extra` attribute.
+    "ESTIMATE_QUERY_COST": False,
     # Allow users to enable ssh tunneling when creating a DB.
     # Users must check whether the DB engine supports SSH Tunnels
     # otherwise enabling this flag won't have any effect on the DB.
@@ -932,16 +937,14 @@ class CeleryConfig:  # pylint: disable=too-few-public-methods
 # query costs before they run. These EXPLAIN queries should have a small
 # timeout.
 SQLLAB_QUERY_COST_ESTIMATE_TIMEOUT = int(timedelta(seconds=10).total_seconds())
-# The feature is off by default, and currently only supported in Presto and Postgres.
-# It also need to be enabled on a per-database basis, by adding the key/value pair
-# `cost_estimate_enabled: true` to the database `extra` attribute.
-ESTIMATE_QUERY_COST = False
+
 # The cost returned by the databases is a relative value; in order to map the cost to
 # a tangible value you need to define a custom formatter that takes into consideration
 # your specific infrastructure. For example, you could analyze queries a posteriori by
 # running EXPLAIN on them, and compute a histogram of relative costs to present the
-# cost as a percentile:
-#
+# cost as a percentile, this step is optional as every db engine spec has its own
+# query cost formatter, but it you wanna customize it you can define it inside the config:
+
 # def postgres_query_cost_formatter(
 #     result: List[Dict[str, Any]]
 # ) -> List[Dict[str, str]]:
@@ -959,9 +962,7 @@ class CeleryConfig:  # pylint: disable=too-few-public-methods
 #
 #     return out
 #
-#  Then on define the formatter on the config:
-#
-# "QUERY_COST_FORMATTERS_BY_ENGINE": {"postgresql": postgres_query_cost_formatter},
+# QUERY_COST_FORMATTERS_BY_ENGINE: {"postgresql": postgres_query_cost_formatter}
 QUERY_COST_FORMATTERS_BY_ENGINE: Dict[
     str, Callable[[List[Dict[str, Any]]], List[Dict[str, Any]]]
 ] = {}

superset/db_engine_specs/bigquery.py

@@ -31,6 +31,7 @@
 from sqlalchemy.sql import sqltypes
 from typing_extensions import TypedDict
 
+from superset import sql_parse
 from superset.constants import PASSWORD_MASK
 from superset.databases.schemas import encrypted_field_properties, EncryptedString
 from superset.databases.utils import make_url_safe
@@ -364,6 +365,97 @@ def df_to_sql(
 
         pandas_gbq.to_gbq(df, **to_gbq_kwargs)
 
+    @classmethod
+    def estimate_query_cost(
+        cls,
+        database: "Database",
+        schema: str,
+        sql: str,
+        source: Optional[utils.QuerySource] = None,
+    ) -> List[Dict[str, Any]]:
+        """
+        Estimate the cost of a multiple statement SQL query.
+
+        :param database: Database instance
+        :param schema: Database schema
+        :param sql: SQL query with possibly multiple statements
+        :param source: Source of the query (eg, "sql_lab")
+        """
+        extra = database.get_extra() or {}
+        if not cls.get_allow_cost_estimate(extra):
+            raise Exception("Database does not support cost estimation")
+
+        parsed_query = sql_parse.ParsedQuery(sql)
+        statements = parsed_query.get_statements()
+        costs = []
+        for statement in statements:
+            processed_statement = cls.process_statement(statement, database)
+
+            costs.append(cls.estimate_statement_cost(processed_statement, database))
+        return costs
+
+    @classmethod
+    def get_allow_cost_estimate(cls, extra: Dict[str, Any]) -> bool:
+        return True
+
+    @classmethod
+    def estimate_statement_cost(cls, statement: str, cursor: Any) -> Dict[str, Any]:
+        try:
+            # pylint: disable=import-outside-toplevel
+            # It's the only way to perfom a dry-run estimate cost
+            from google.cloud import bigquery
+            from google.oauth2 import service_account
+        except ImportError as ex:
+            raise Exception(
+                "Could not import libraries `pygibquery` or `google.oauth2`, which are "
+                "required to be installed in your environment in order "
+                "to upload data to BigQuery"
+            ) from ex
+
+        with cls.get_engine(cursor) as engine:
+            creds = engine.dialect.credentials_info
+
+        creds = service_account.Credentials.from_service_account_info(creds)
+        client = bigquery.Client(credentials=creds)
+        job_config = bigquery.QueryJobConfig(dry_run=True)
+
+        query_job = client.query(
+            statement,
+            job_config=job_config,
+        )  # Make an API request.
+
+        # Format Bytes.
+        # TODO: Humanize in case more db engine specs need to be added,
+        # this should be made a function outside this scope.
+        byte_division = 1024
+        if hasattr(query_job, "total_bytes_processed"):
+            query_bytes_processed = query_job.total_bytes_processed
+            if query_bytes_processed // byte_division == 0:
+                byte_type = "B"
+                total_bytes_processed = query_bytes_processed
+            elif query_bytes_processed // (byte_division**2) == 0:
+                byte_type = "KB"
+                total_bytes_processed = round(query_bytes_processed / byte_division, 2)
+            elif query_bytes_processed // (byte_division**3) == 0:
+                byte_type = "MB"
+                total_bytes_processed = round(
+                    query_bytes_processed / (byte_division**2), 2
+                )
+            else:
+                byte_type = "GB"
+                total_bytes_processed = round(
+                    query_bytes_processed / (byte_division**3), 2
+                )
+
+            return {f"{byte_type} Processed": total_bytes_processed}
+        return {}
+
+    @classmethod
+    def query_cost_formatter(
+        cls, raw_cost: List[Dict[str, Any]]
+    ) -> List[Dict[str, str]]:
+        return [{k: str(v) for k, v in row.items()} for row in raw_cost]
+
     @classmethod
     def build_sqlalchemy_uri(
         cls,

superset/translations/de/LC_MESSAGES/messages.json

@@ -1963,7 +1963,7 @@
       "Font size for the smallest value in the list": [
         "Schriftgröße für den kleinsten Wert in der Liste"
       ],
-      "For Presto and Postgres, shows a button to compute cost before running a query.": [
+      "For Bigquery, Presto and Postgres, shows a button to compute cost before running a query.": [
         "Für Presto und Postgres wird ein Buttons angezeigt, um Kosten vor dem Ausführen einer Abfrage zu schätzen."
       ],
       "For regular filters, these are the roles this filter will be applied to. For base filters, these are the roles that the filter DOES NOT apply to, e.g. Admin if admin should see all data.": [

superset/translations/de/LC_MESSAGES/messages.po

@@ -6114,7 +6114,7 @@ msgstr "Schriftgröße für den kleinsten Wert in der Liste"
 
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr ""
 "Für Presto und Postgres wird ein Buttons angezeigt, um Kosten vor dem "

superset/translations/en/LC_MESSAGES/messages.po

@@ -5693,7 +5693,7 @@ msgstr ""
 
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr ""
 

superset/translations/es/LC_MESSAGES/messages.po

@@ -5975,7 +5975,7 @@ msgstr ""
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 #, fuzzy
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr "Estimar el costo antes de ejecutar una consulta"
 

superset/translations/fr/LC_MESSAGES/messages.json

@@ -3419,7 +3419,7 @@
       "Enable query cost estimation": [
         "Activer l'estimation du coût de la requête"
       ],
-      "For Presto and Postgres, shows a button to compute cost before running a query.": [
+      "For Bigquery, Presto and Postgres, shows a button to compute cost before running a query.": [
         "Pour Presto et Postgres, affiche un bouton pour calculer le coût avant d'exécuter une requête."
       ],
       "Allow this database to be explored": [

superset/translations/fr/LC_MESSAGES/messages.po

@@ -6118,7 +6118,7 @@ msgstr ""
 
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr ""
 "Pour Presto et Postgres, affiche un bouton pour calculer le coût avant "

superset/translations/it/LC_MESSAGES/messages.po

@@ -5848,7 +5848,7 @@ msgstr ""
 
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr ""
 

superset/translations/ja/LC_MESSAGES/messages.po

@@ -5833,7 +5833,7 @@ msgstr ""
 
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr ""
 

superset/translations/ko/LC_MESSAGES/messages.po

@@ -5798,7 +5798,7 @@ msgstr ""
 
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr ""
 

superset/translations/messages.pot

@@ -5698,7 +5698,7 @@ msgstr ""
 
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr ""
 

superset/translations/nl/LC_MESSAGES/messages.json

@@ -4487,7 +4487,7 @@
         "Sta manipulatie van de database toe met niet-SELECT statements zoals UPDATE, DELETE, CREATE, enz."
       ],
       "Enable query cost estimation": [""],
-      "For Presto and Postgres, shows a button to compute cost before running a query.": [
+      "For Bigquery, Presto and Postgres, shows a button to compute cost before running a query.": [
         ""
       ],
       "Allow this database to be explored": [""],

superset/translations/nl/LC_MESSAGES/messages.po

@@ -15015,7 +15015,7 @@ msgstr ""
 
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr ""
 

superset/translations/pt_BR/LC_MESSAGES/messages.po

@@ -6115,7 +6115,7 @@ msgstr ""
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 #, fuzzy
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr "Estima o custo antes de executar uma consulta"
 

superset/translations/ru/LC_MESSAGES/messages.po

@@ -6040,7 +6040,7 @@ msgstr ""
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 #, fuzzy
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr "Спрогнозировать время до выполнения запроса"
 

superset/translations/sk/LC_MESSAGES/messages.po

@@ -5710,7 +5710,7 @@ msgstr ""
 
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr ""
 

superset/translations/sl/LC_MESSAGES/messages.json

@@ -3453,17 +3453,8 @@
       "Are you sure you want to overwrite this dataset?": [
         "Ali ste prepričani, da želite prepisati podatkovni set?"
       ],
-      "Undefined": ["Ni definirano"],
-      "Save": ["Shrani"],
-      "Save as": ["Shrani kot"],
-      "Save query": ["Shrani poizvedbo"],
-      "Update": ["Posodobi"],
-      "Label for your query": ["Ime vaše poizvedbe"],
-      "Write a description for your query": ["Dodajte opis vaše poizvedbe"],
-      "Schedule query": ["Urnik poizvedb"],
-      "Schedule": ["Urnik"],
-      "There was an error with your request": [
-        "Pri zahtevi je prišlo do napake"
+      "For Bigquery, Presto and Postgres, shows a button to compute cost before running a query.": [
+        "Za Presto in Postgres prikaže gumb za izračun potratnosti pred zagonom poizvedbe."
       ],
       "Please save the query to enable sharing": [
         "Shranite poizvedbo za deljenje"
@@ -5138,7 +5129,7 @@
       "Enable query cost estimation": [
         "Omogoči ocenjevanje potratnosti poizvedbe"
       ],
-      "For Presto and Postgres, shows a button to compute cost before running a query.": [
+      "For Bigquery, Presto and Postgres, shows a button to compute cost before running a query.": [
         "Za Presto in Postgres prikaže gumb za izračun potratnosti pred zagonom poizvedbe."
       ],
       "Allow this database to be explored": [

superset/translations/sl/LC_MESSAGES/messages.po

@@ -6696,9 +6696,8 @@ msgstr "Ciljno razmerje za razdelke drevesnega grafikona."
 
 #: superset-frontend/plugins/legacy-plugin-chart-treemap/src/index.js:31
 msgid ""
-"Shows the composition of a dataset by segmenting a given rectangle as smaller "
-"rectangles with areas proportional to their value or contribution to the whole. "
-"Those rectangles may also, in turn, be further segmented hierarchically."
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
+"query."
 msgstr ""
 "Prikaže zgradbo podatkovnega seta na podlagi segmentacije danega pravokotnika na "
 "manjše pravokotnike, pri čemer je ploščina sorazmerna vrednostim oz. deležem. "
@@ -15788,7 +15787,7 @@ msgstr "Omogoči ocenjevanje potratnosti poizvedbe"
 
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a query."
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a query."
 msgstr ""
 "Za Presto in Postgres prikaže gumb za izračun potratnosti pred zagonom poizvedbe."
 

superset/translations/zh/LC_MESSAGES/messages.po

@@ -5924,7 +5924,7 @@ msgstr "列表中最小值的字体大小"
 #: superset-frontend/src/views/CRUD/data/database/DatabaseModal/ExtraOptions.tsx:179
 #, fuzzy
 msgid ""
-"For Presto and Postgres, shows a button to compute cost before running a "
+"For Bigquery, Presto and Postgres, shows a button to compute cost before running a "
 "query."
 msgstr "在运行查询之前计算执行计划"