ingestr 0.13.51__py3-none-any.whl → 0.13.52__py3-none-any.whl

ingestr/main.py

@@ -58,6 +58,7 @@ class LoaderFileFormat(str, Enum):
 
 
 class SqlBackend(str, Enum):
+    default = "default"
     sqlalchemy = "sqlalchemy"
     pyarrow = "pyarrow"
     connectorx = "connectorx"
@@ -187,7 +188,7 @@ def ingest(
             help="The SQL backend to use",
             envvar=["SQL_BACKEND", "INGESTR_SQL_BACKEND"],
         ),
-    ] = SqlBackend.pyarrow,  # type: ignore
+    ] = SqlBackend.default,  # type: ignore
     loader_file_format: Annotated[
         Optional[LoaderFileFormat],
         typer.Option(
@@ -289,7 +290,11 @@ def ingest(
     from ingestr.src.collector.spinner import SpinnerCollector
     from ingestr.src.destinations import AthenaDestination
     from ingestr.src.factory import SourceDestinationFactory
-    from ingestr.src.filters import cast_set_to_list, handle_mysql_empty_dates
+    from ingestr.src.filters import (
+        cast_set_to_list,
+        cast_spanner_types,
+        handle_mysql_empty_dates,
+    )
     from ingestr.src.sources import MongoDbSource
 
     def report_errors(run_info: LoadInfo):
@@ -517,6 +522,15 @@ def ingest(
             if interval_end:
                 interval_end = interval_end.date()  # type: ignore
 
+        if factory.source_scheme.startswith("spanner"):
+            # we tend to use the 'pyarrow' backend in general, however, it has issues with JSON objects, so we override it to 'sqlalchemy' for Spanner.
+            if sql_backend.value == SqlBackend.default:
+                sql_backend = SqlBackend.sqlalchemy
+
+        # this allows us to identify the cases where the user does not have a preference, so that for some sources we can override it.
+        if sql_backend == SqlBackend.default:
+            sql_backend = SqlBackend.pyarrow
+
         dlt_source = source.dlt_source(
             uri=source_uri,
             table=source_table,
@@ -535,6 +549,9 @@ def ingest(
         if factory.source_scheme.startswith("mysql"):
             resource.for_each(dlt_source, lambda x: x.add_map(handle_mysql_empty_dates))
 
+        if factory.source_scheme.startswith("spanner"):
+            resource.for_each(dlt_source, lambda x: x.add_map(cast_spanner_types))
+
         if yield_limit:
             resource.for_each(dlt_source, lambda x: x.add_limit(yield_limit))
 

ingestr/src/buildinfo.py

@@ -1 +1 @@
-version = "v0.13.51"
+version = "v0.13.52"

ingestr/src/destinations.py

@@ -468,3 +468,15 @@ class S3Destination:
 
     def post_load(self) -> None:
         pass
+
+
+class SqliteDestination(GenericSqlDestination):
+    def dlt_dest(self, uri: str, **kwargs):
+        return dlt.destinations.sqlalchemy(credentials=uri)
+
+    def dlt_run_params(self, uri: str, table: str, **kwargs):
+        return {
+            #https://dlthub.com/docs/dlt-ecosystem/destinations/sqlalchemy#dataset-files
+            "dataset_name": "main",
+            "table_name": table,
+        }

ingestr/src/factory.py

@@ -15,6 +15,7 @@ from ingestr.src.destinations import (
     RedshiftDestination,
     S3Destination,
     SnowflakeDestination,
+    SqliteDestination,
     SynapseDestination,
 )
 from ingestr.src.sources import (
@@ -54,12 +55,12 @@ from ingestr.src.sources import (
     SalesforceSource,
     ShopifySource,
     SlackSource,
+    SmartsheetSource,
     SolidgateSource,
     SqlSource,
     StripeAnalyticsSource,
     TikTokSource,
     ZendeskSource,
-    SmartsheetSource,
 )
 
 SQL_SOURCE_SCHEMES = [
@@ -182,6 +183,7 @@ class SourceDestinationFactory:
         "clickhouse+native": ClickhouseDestination,
         "clickhouse": ClickhouseDestination,
         "s3": S3Destination,
+        "sqlite": SqliteDestination,
     }
 
     def __init__(self, source_uri: str, destination_uri: str):

ingestr/src/filters.py

@@ -7,6 +7,20 @@ def cast_set_to_list(row):
     return row
 
 
+def cast_spanner_types(row):
+    if not isinstance(row, dict):
+        return row
+
+    from google.cloud.spanner_v1.data_types import JsonObject
+
+    for key in row.keys():
+        if isinstance(row[key], JsonObject):
+            import json
+
+            row[key] = json.loads(row[key].serialize())
+    return row
+
+
 def handle_mysql_empty_dates(row):
     # MySQL returns empty dates as 0000-00-00, which is not a valid date, we handle them here.
     if not isinstance(row, dict):

ingestr/src/solidgate/__init__.py

@@ -43,7 +43,7 @@ def solidgate_source(
         yield solidgate_client.fetch_data(path, date_from=start_dt, date_to=end_dt)
 
     @dlt.resource(
-        name="apm-orders",
+        name="apm_orders",
         write_disposition="merge",
         primary_key="order_id",
         columns={
@@ -69,7 +69,7 @@ def solidgate_source(
         yield solidgate_client.fetch_data(path, date_from=start_dt, date_to=end_dt)
 
     @dlt.resource(
-        name="card-orders",
+        name="card_orders",
         write_disposition="merge",
         primary_key="order_id",
         columns={
@@ -95,7 +95,7 @@ def solidgate_source(
         yield solidgate_client.fetch_data(path, date_from=start_dt, date_to=end_dt)
 
     @dlt.resource(
-        name="financial-entries",
+        name="financial_entries",
         write_disposition="merge",
         primary_key="id",
         columns={

{ingestr-0.13.51.dist-info → ingestr-0.13.52.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ingestr
-Version: 0.13.51
+Version: 0.13.52
 Summary: ingestr is a command-line application that ingests data from various sources and stores them in any database.
 Project-URL: Homepage, https://github.com/bruin-data/ingestr
 Project-URL: Issues, https://github.com/bruin-data/ingestr/issues

{ingestr-0.13.51.dist-info → ingestr-0.13.52.dist-info}/RECORD

@@ -1,12 +1,12 @@
 ingestr/conftest.py,sha256=Q03FIJIZpLBbpj55cfCHIKEjc1FCvWJhMF2cidUJKQU,1748
-ingestr/main.py,sha256=rHxHQAbd0ccW2e2kQSWlv7-5qcc2ZB6Eh3vyjm4Nzns,25550
+ingestr/main.py,sha256=GkC1hdq8AVGrvolc95zMfjmibI95p2pmFkbgCOVf-Og,26311
 ingestr/src/.gitignore,sha256=8cX1AZTSI0TcdZFGTmS_oyBjpfCzhOEt0DdAo2dFIY8,203
 ingestr/src/blob.py,sha256=onMe5ZHxPXTdcB_s2oGNdMo-XQJ3ajwOsWE9eSTGFmc,1495
-ingestr/src/buildinfo.py,sha256=ZgugQZPIlr-_7kRH5qIupQ5e48JTOHSlRWEsmNx4wXw,21
-ingestr/src/destinations.py,sha256=41Bj1UgxR8a2KcZWqtGw74AKZKnSBrueQRnBdrf3c-A,16003
+ingestr/src/buildinfo.py,sha256=qcKqSPI861BBapk7rbcYkc9M_18QQnT44YM1Zr_Sg80,21
+ingestr/src/destinations.py,sha256=axSm0GLM_bKQmWYb74fnHebNktWgtfAyyqdHV-zBOsA,16405
 ingestr/src/errors.py,sha256=Ufs4_DfE77_E3vnA1fOQdi6cmuLVNm7_SbFLkL1XPGk,686
-ingestr/src/factory.py,sha256=JeK0M8C05-_KxxIhtevs-MJGLlngXUQupd-Pm4Fg904,5803
-ingestr/src/filters.py,sha256=C-_TIVkF_cxZBgG-Run2Oyn0TAhJgA8IWXZ-OPY3uek,1136
+ingestr/src/factory.py,sha256=5yrg-XkaixuCkiTz3B7mraE8LaANXtzItenbx8TdPrE,5863
+ingestr/src/filters.py,sha256=LLecXe9QkLFkFLUZ92OXNdcANr1a8edDxrflc2ko_KA,1452
 ingestr/src/http_client.py,sha256=bxqsk6nJNXCo-79gW04B53DQO-yr25vaSsqP0AKtjx4,732
 ingestr/src/loader.py,sha256=9NaWAyfkXdqAZSS-N72Iwo36Lbx4PyqIfaaH1dNdkFs,1712
 ingestr/src/partition.py,sha256=BrIP6wFJvyR7Nus_3ElnfxknUXeCipK_E_bB8kZowfc,969
@@ -109,7 +109,7 @@ ingestr/src/slack/__init__.py,sha256=pyDukxcilqTAe_bBzfWJ8Vxi83S-XEdEFBH2pEgILrM
 ingestr/src/slack/helpers.py,sha256=08TLK7vhFvH_uekdLVOLF3bTDe1zgH0QxHObXHzk1a8,6545
 ingestr/src/slack/settings.py,sha256=NhKn4y1zokEa5EmIZ05wtj_-I0GOASXZ5V81M1zXCtY,457
 ingestr/src/smartsheets/__init__.py,sha256=pdzSV7rA0XYD5Xa1u4zb6vziy5iFXIQNROkpJ9oYas0,1623
-ingestr/src/solidgate/__init__.py,sha256=DZYQ4M3Cc7AIbdQcNQm_6yX2whnFhE-iM10-ACJ3W3A,3626
+ingestr/src/solidgate/__init__.py,sha256=JdaXvAu5QGuf9-FY294vwCQCEmfrqIld9oqbzqCJS3g,3626
 ingestr/src/solidgate/helpers.py,sha256=oePEc9nnvmN3IaKrfJCvyKL79xdGM0-gRTN3-8tY4Fc,4952
 ingestr/src/sql_database/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ingestr/src/sql_database/callbacks.py,sha256=sEFFmXxAURY3yeBjnawigDtq9LBCvi8HFqG4kLd7tMU,2002
@@ -135,8 +135,8 @@ ingestr/testdata/merge_expected.csv,sha256=DReHqWGnQMsf2PBv_Q2pfjsgvikYFnf1zYcQZ
 ingestr/testdata/merge_part1.csv,sha256=Pw8Z9IDKcNU0qQHx1z6BUf4rF_-SxKGFOvymCt4OY9I,185
 ingestr/testdata/merge_part2.csv,sha256=T_GiWxA81SN63_tMOIuemcvboEFeAmbKc7xRXvL9esw,287
 ingestr/tests/unit/test_smartsheets.py,sha256=eiC2CCO4iNJcuN36ONvqmEDryCA1bA1REpayHpu42lk,5058
-ingestr-0.13.51.dist-info/METADATA,sha256=danPRyBhqlvyOwtw8kBmBh5FEY_xH6upsy5GrsFaMg4,13983
-ingestr-0.13.51.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-ingestr-0.13.51.dist-info/entry_points.txt,sha256=oPJy0KBnPWYjDtP1k8qwAihcTLHSZokSQvRAw_wtfJM,46
-ingestr-0.13.51.dist-info/licenses/LICENSE.md,sha256=cW8wIhn8HFE-KLStDF9jHQ1O_ARWP3kTpk_-eOccL24,1075
-ingestr-0.13.51.dist-info/RECORD,,
+ingestr-0.13.52.dist-info/METADATA,sha256=jfSWJ6mE6vPz-5XsIIjqdLxd45-OrOUKbhbwJqiZlSc,13983
+ingestr-0.13.52.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+ingestr-0.13.52.dist-info/entry_points.txt,sha256=oPJy0KBnPWYjDtP1k8qwAihcTLHSZokSQvRAw_wtfJM,46
+ingestr-0.13.52.dist-info/licenses/LICENSE.md,sha256=cW8wIhn8HFE-KLStDF9jHQ1O_ARWP3kTpk_-eOccL24,1075
+ingestr-0.13.52.dist-info/RECORD,,