ingestr 0.13.36__py3-none-any.whl → 0.13.38__py3-none-any.whl

ingestr/main.py

@@ -485,6 +485,9 @@ def ingest(
         print(
             f"[bold yellow]  Primary Key:[/bold yellow] {primary_key if primary_key else 'None'}"
         )
+        print(
+            f"[bold yellow]  Pipeline ID:[/bold yellow] {m.hexdigest()}"
+        )
         print()
 
         if not yes:

ingestr/src/buildinfo.py

@@ -1 +1 @@
-version = "v0.13.36"
+version = "v0.13.38"

ingestr/src/factory.py

@@ -54,6 +54,7 @@ from ingestr.src.sources import (
     TikTokSource,
     ZendeskSource,
     FreshdeskSource,
+    PhantombusterSource,
 )
 
 SQL_SOURCE_SCHEMES = [
@@ -150,6 +151,7 @@ class SourceDestinationFactory:
         "pipedrive": PipedriveSource,
         "frankfurter": FrankfurterSource,
         "freshdesk": FreshdeskSource,
+        "phantombuster": PhantombusterSource,
     }
     destinations: Dict[str, Type[DestinationProtocol]] = {
         "bigquery": BigQueryDestination,

ingestr/src/frankfurter/__init__.py

@@ -1,4 +1,4 @@
-from typing import Any, Iterator
+from typing import Any, Iterator, Optional
 
 import dlt
 from dlt.common.pendulum import pendulum
@@ -15,13 +15,13 @@ from ingestr.src.frankfurter.helpers import get_path_with_retry
 def frankfurter_source(
     start_date: TAnyDateTime,
     end_date: TAnyDateTime,
+    base_currency: str,
 ) -> Any:
     """
     A dlt source for the frankfurter.dev API. It groups several resources (in this case frankfurter.dev API endpoints) containing
     various types of data: currencies, latest rates, historical rates.
     """
     date_time = dlt.sources.incremental(
-
         "date",
         initial_value=start_date,
         end_value=end_date,
@@ -31,9 +31,10 @@ def frankfurter_source(
 
     return (
         currencies(),
-        latest(),
-        exchange_rates(start_date=date_time, end_date=end_date),
-
+        latest(base_currency=base_currency),
+        exchange_rates(
+            start_date=date_time, end_date=end_date, base_currency=base_currency
+        ),
     )
 
 
@@ -61,29 +62,33 @@ def currencies() -> Iterator[dict]:
         "date": {"data_type": "text"},
         "currency_code": {"data_type": "text"},
         "rate": {"data_type": "double"},
+        "base_currency": {"data_type": "text"},
     },
-    primary_key=["date", "currency_code"],  # Composite primary key
+    primary_key=["date", "currency_code", "base_currency"],
 )
-def latest() -> Iterator[dict]:
+def latest(base_currency: Optional[str] = "") -> Iterator[dict]:
     """
     Fetches the latest exchange rates and yields them as rows.
     """
     # Base URL
     url = "latest?"
 
+    if base_currency:
+        url += f"base={base_currency}"
+
     # Fetch data
     data = get_path_with_retry(url)
 
     # Extract rates and base currency
     rates = data["rates"]
-
     date = pendulum.parse(data["date"])
 
-    # Add the base currency (EUR) with a rate of 1.0
+    # Add the base currency with a rate of 1.0
     yield {
         "date": date,
-        "currency_code": "EUR",
+        "currency_code": base_currency,
         "rate": 1.0,
+        "base_currency": base_currency,
     }
 
     # Add all currencies and their rates
@@ -92,6 +97,7 @@ def latest() -> Iterator[dict]:
             "date": date,
             "currency_code": currency_code,
             "rate": rate,
+            "base_currency": base_currency,
         }
 
 
@@ -101,12 +107,14 @@ def latest() -> Iterator[dict]:
         "date": {"data_type": "text"},
         "currency_code": {"data_type": "text"},
         "rate": {"data_type": "double"},
+        "base_currency": {"data_type": "text"},
     },
-    primary_key=("date", "currency_code"),  # Composite primary key
+    primary_key=("date", "currency_code", "base_currency"),
 )
 def exchange_rates(
     end_date: TAnyDateTime,
     start_date: dlt.sources.incremental[TAnyDateTime] = dlt.sources.incremental("date"),
+    base_currency: Optional[str] = "",
 ) -> Iterator[dict]:
     """
     Fetches exchange rates for a specified date range.
@@ -124,6 +132,9 @@ def exchange_rates(
     # Compose the URL
     url = f"{start_date_str}..{end_date_str}?"
 
+    if base_currency:
+        url += f"base={base_currency}"
+
     # Fetch data from the API
     data = get_path_with_retry(url)
 
@@ -137,8 +148,9 @@ def exchange_rates(
         # Add the base currency with a rate of 1.0
         yield {
             "date": formatted_date,
-            "currency_code": "EUR",
+            "currency_code": base_currency,
             "rate": 1.0,
+            "base_currency": base_currency,
         }
 
         # Add all other currencies and their rates
@@ -147,4 +159,5 @@ def exchange_rates(
                 "date": formatted_date,
                 "currency_code": currency_code,
                 "rate": rate,
+                "base_currency": base_currency,
             }

ingestr/src/frankfurter/helpers.py

@@ -30,3 +30,19 @@ def validate_dates(start_date: datetime, end_date: datetime) -> None:
     # Check if start_date is before end_date
     if start_date > end_date:
         raise ValueError("Interval-end cannot be before interval-start.")
+
+
+def validate_currency(currency_code: str) -> bool:
+    url = "https://api.frankfurter.dev/v1/currencies"
+
+    response = requests.get(url, timeout=5)
+    currencies = response.json()
+
+    if currency_code.upper() in currencies:
+        return True
+    else:
+        supported_currencies = list(currencies.keys())
+        print(
+            f"Invalid base currency '{currency_code}'. Supported currencies are: {supported_currencies}"
+        )
+        return False

ingestr/src/phantombuster/__init__.py

@@ -0,0 +1,60 @@
+from typing import Iterable, Optional
+
+import dlt
+import pendulum
+import requests
+from dlt.common.typing import TDataItem, TAnyDateTime
+from dlt.sources import DltResource
+from dlt.sources.helpers.requests import Client
+
+
+from ingestr.src.phantombuster.client import PhantombusterClient
+
+
+def retry_on_limit(
+    response: Optional[requests.Response], exception: Optional[BaseException]
+) -> bool:
+    if response is not None and response.status_code == 429:
+        return True
+    return False
+
+
+def create_client() -> requests.Session:
+    return Client(
+        raise_for_status=False,
+        retry_condition=retry_on_limit,
+        request_max_attempts=12,
+        request_backoff_factor=2,
+    ).session
+
+@dlt.source(max_table_nesting=0)
+def phantombuster_source(api_key: str, agent_id: str, start_date: TAnyDateTime, end_date: TAnyDateTime | None) -> Iterable[DltResource]:
+    client = PhantombusterClient(api_key)
+    @dlt.resource(
+        write_disposition="merge",
+        primary_key="container_id",
+        columns={
+            "partition_dt": {"data_type": "date", "partition": True},
+        },
+    )
+    def completed_phantoms(
+        dateTime=(
+            dlt.sources.incremental(
+                "ended_at",
+                initial_value=start_date,
+                end_value=end_date,
+                range_start="closed",
+                range_end="closed",
+            )
+        ),
+    ) -> Iterable[TDataItem]:
+        if dateTime.end_value is None:
+            end_dt = pendulum.now(tz="UTC")
+        else:
+            end_dt = dateTime.end_value
+
+        start_dt = dateTime.last_value
+
+        yield client.fetch_containers_result(create_client(), agent_id, start_date=start_dt, end_date=end_dt)
+
+    return completed_phantoms

ingestr/src/phantombuster/client.py

@@ -0,0 +1,70 @@
+from typing import Union
+
+import pendulum
+import requests
+
+
+class PhantombusterClient:
+    def __init__(self, api_key: str):
+        self.api_key = api_key
+
+    def _get_headers(self):
+        return {
+            "X-Phantombuster-Key-1": self.api_key,
+            "accept": "application/json",
+        }
+
+    def fetch_containers_result(self, session: requests.Session, agent_id: str, start_date: pendulum.DateTime, end_date: pendulum.DateTime):
+        url = "https://api.phantombuster.com/api/v2/containers/fetch-all/"
+        before_ended_at = None
+        limit = 100
+
+        started_at = start_date.int_timestamp * 1000 + int(start_date.microsecond / 1000)
+        ended_at = end_date.int_timestamp * 1000 + int(end_date.microsecond / 1000)
+       
+        while True:
+            params: dict[str, Union[str, int, float, bytes, None]] = {
+                "agentId": agent_id,
+                "limit": limit,
+                "mode": "finalized",
+            }
+
+            if before_ended_at:
+                params["beforeEndedAt"] = before_ended_at
+
+            response = session.get(url=url, headers=self._get_headers(), params=params)
+            data = response.json()
+            containers = data.get("containers", [])
+
+            for container in containers:
+                container_ended_at = container.get("endedAt")
+
+                if before_ended_at is None or before_ended_at > container_ended_at:
+                        before_ended_at = container_ended_at
+                
+                if container_ended_at < started_at or container_ended_at > ended_at:
+                    continue
+              
+                try:
+                    result = self.fetch_result_object(session, container["id"])
+                    partition_dt = pendulum.from_timestamp(container_ended_at / 1000, tz="UTC").date()
+                    container_ended_at_datetime = pendulum.from_timestamp(container_ended_at / 1000, tz="UTC")
+                    row = {"container_id": container["id"],"container": container, "result": result, "partition_dt": partition_dt, "ended_at": container_ended_at_datetime}
+                    yield row
+                    
+                except requests.RequestException as e:
+                    print(f"Error fetching result for container {container['id']}: {e}")
+         
+            if data["maxLimitReached"] is False:
+                break
+ 
+
+    def fetch_result_object(self, session: requests.Session, container_id: str):
+        result_url = (
+            "https://api.phantombuster.com/api/v2/containers/fetch-result-object"
+        )
+        params = {"id": container_id}
+        response = session.get(result_url, headers=self._get_headers(), params=params)
+        response.raise_for_status()
+
+        return response.json()

ingestr/src/sources.py

@@ -2180,6 +2180,18 @@ class FrankfurterSource:
                 "Frankfurter takes care of incrementality on its own, you should not provide incremental_key"
             )
 
+        from ingestr.src.frankfurter import frankfurter_source
+        from ingestr.src.frankfurter.helpers import validate_currency, validate_dates
+
+        parsed_uri = urlparse(uri)
+        source_params = parse_qs(parsed_uri.query)
+        base_currency = source_params.get("base", [None])[0]
+
+        if not base_currency:
+            base_currency = "USD"
+
+        validate_currency(base_currency)
+
         if kwargs.get("interval_start"):
             start_date = ensure_pendulum_datetime(str(kwargs.get("interval_start")))
             if kwargs.get("interval_end"):
@@ -2190,21 +2202,20 @@ class FrankfurterSource:
             start_date = pendulum.now()
             end_date = pendulum.now()
 
-        from ingestr.src.frankfurter import frankfurter_source
-        from ingestr.src.frankfurter.helpers import validate_dates
-
         validate_dates(start_date=start_date, end_date=end_date)
 
         src = frankfurter_source(
             start_date=start_date,
             end_date=end_date,
+            base_currency=base_currency,
         )
 
         if table not in src.resources:
             raise UnsupportedResourceError(table, "Frankfurter")
 
         return src.with_resources(table)
-    
+
+
 class FreshdeskSource:
      # freshdesk://domain?api_key=<api_key>
     def handles_incrementality(self) -> bool:
@@ -2231,4 +2242,40 @@ class FreshdeskSource:
         
         from ingestr.src.freshdesk import freshdesk_source
         return freshdesk_source(api_secret_key=api_key[0], domain=domain).with_resources(table)
-   
+
+class PhantombusterSource:
+    def handles_incrementality(self) -> bool:
+        return True
+    
+    def dlt_source(self, uri: str, table: str, **kwargs):
+        #phantombuster://?api_key=<api_key>
+        #source table = phantom_results:agent_id
+        parsed_uri = urlparse(uri)
+        params = parse_qs(parsed_uri.query)
+        api_key = params.get("api_key")
+        if api_key is None:
+            raise MissingValueError("api_key", "Phantombuster")
+        
+        table_fields = table.replace(" ", "").split(":")
+        table_name = table_fields[0]
+        
+        agent_id = table_fields[1] if len(table_fields) > 1 else None
+        
+        if table_name not in ["completed_phantoms"]:
+            raise UnsupportedResourceError(table_name, "Phantombuster")
+        
+        if not agent_id:
+            raise MissingValueError("agent_id", "Phantombuster")
+        
+        start_date = kwargs.get("interval_start")
+        if start_date is None:
+            start_date = ensure_pendulum_datetime("2018-01-01").in_tz("UTC")
+        else:
+            start_date = ensure_pendulum_datetime(start_date).in_tz("UTC")
+
+        end_date = kwargs.get("interval_end")
+        if end_date is not None:
+            end_date = ensure_pendulum_datetime(end_date).in_tz("UTC")
+        
+        from ingestr.src.phantombuster import phantombuster_source
+        return phantombuster_source(api_key=api_key[0], agent_id=agent_id, start_date=start_date, end_date=end_date).with_resources(table_name)

{ingestr-0.13.36.dist-info → ingestr-0.13.38.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ingestr
-Version: 0.13.36
+Version: 0.13.38
 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
@@ -74,7 +74,7 @@ Requires-Dist: google-cloud-storage==3.1.0
 Requires-Dist: google-crc32c==1.6.0
 Requires-Dist: google-resumable-media==2.7.2
 Requires-Dist: googleapis-common-protos==1.69.0
-Requires-Dist: greenlet==3.2.1
+Requires-Dist: greenlet==3.2.2
 Requires-Dist: grpcio-status==1.62.3
 Requires-Dist: grpcio==1.70.0
 Requires-Dist: hdbcli==2.23.27

{ingestr-0.13.36.dist-info → ingestr-0.13.38.dist-info}/RECORD

@@ -1,16 +1,16 @@
 ingestr/conftest.py,sha256=Q03FIJIZpLBbpj55cfCHIKEjc1FCvWJhMF2cidUJKQU,1748
-ingestr/main.py,sha256=mRlGSqi2sHcZ2AKlwn5MqoMvFxXlSjcZxmPJr76rmRk,25187
+ingestr/main.py,sha256=QHLjpCItCgL5XHcTezBOp6Vdy_VHAPciRjqy63ZIK2s,25285
 ingestr/src/.gitignore,sha256=8cX1AZTSI0TcdZFGTmS_oyBjpfCzhOEt0DdAo2dFIY8,203
 ingestr/src/blob.py,sha256=onMe5ZHxPXTdcB_s2oGNdMo-XQJ3ajwOsWE9eSTGFmc,1495
-ingestr/src/buildinfo.py,sha256=abX1HXd_dkzG2hkJg7JdFGSvgjGi72VrEucHcTxIziA,21
+ingestr/src/buildinfo.py,sha256=iW7GVGJjTCNtuINLzbFuur_d0ALxf_JmZSmuwD9yZ2Y,21
 ingestr/src/destinations.py,sha256=Z79f01BSmEaXnQno2IQVt4Th4dmD-BiOQXlibZJ5sTw,13180
 ingestr/src/errors.py,sha256=Ufs4_DfE77_E3vnA1fOQdi6cmuLVNm7_SbFLkL1XPGk,686
-ingestr/src/factory.py,sha256=M0FAes6KsvqCzuTnUBcxc6DF7UVO51IlrFxy2VDpbkQ,5392
+ingestr/src/factory.py,sha256=Mm_Be60PFO4mUIeJLBMDVU_uyH0IeCiZ1dtNDFiDFSo,5463
 ingestr/src/filters.py,sha256=C-_TIVkF_cxZBgG-Run2Oyn0TAhJgA8IWXZ-OPY3uek,1136
 ingestr/src/loader.py,sha256=9NaWAyfkXdqAZSS-N72Iwo36Lbx4PyqIfaaH1dNdkFs,1712
 ingestr/src/partition.py,sha256=BrIP6wFJvyR7Nus_3ElnfxknUXeCipK_E_bB8kZowfc,969
 ingestr/src/resource.py,sha256=XG-sbBapFVEM7OhHQFQRTdTLlh-mHB-N4V1t8F8Tsww,543
-ingestr/src/sources.py,sha256=YpaUS5Ui-YXeZYLETPAj60WhU5fWI_lP6jVA0w6J6qo,77250
+ingestr/src/sources.py,sha256=q9FJ3vRaeTjbL5sgau_GHPxbn5LfSI4t5TBwycVdwSc,79060
 ingestr/src/table_definition.py,sha256=REbAbqdlmUMUuRh8nEQRreWjPVOQ5ZcfqGkScKdCrmk,390
 ingestr/src/time.py,sha256=H_Fk2J4ShXyUM-EMY7MqCLZQhlnZMZvO952bmZPc4yE,254
 ingestr/src/version.py,sha256=J_2xgZ0mKlvuHcjdKCx2nlioneLH0I47JiU_Slr_Nwc,189
@@ -42,8 +42,8 @@ ingestr/src/facebook_ads/settings.py,sha256=1IxZeP_4rN3IBvAncNHOoqpzAirx0Hz-MUK_
 ingestr/src/filesystem/__init__.py,sha256=zkIwbRr0ir0EUdniI25p2zGiVc-7M9EmR351AjNb0eA,4163
 ingestr/src/filesystem/helpers.py,sha256=bg0muSHZr3hMa8H4jN2-LGWzI-SUoKlQNiWJ74-YYms,3211
 ingestr/src/filesystem/readers.py,sha256=a0fKkaRpnAOGsXI3EBNYZa7x6tlmAOsgRzb883StY30,3987
-ingestr/src/frankfurter/__init__.py,sha256=sjxfq377-lryuFC3JswcbHBRoBjLnGLKNRTwBpDZyLw,4403
-ingestr/src/frankfurter/helpers.py,sha256=wqm087QVPcyTuMl6yj_Pl1wcuqElwcBMPz3P4773wcM,979
+ingestr/src/frankfurter/__init__.py,sha256=oVi4BiOxPRyckEVrBNunyMAHulPyMgyGRwBbhn-Xz6M,4987
+ingestr/src/frankfurter/helpers.py,sha256=SyrkRTDqvKdQxRHTV5kcSeVG3FEnaK5zxHyNyqtumZ0,1445
 ingestr/src/freshdesk/__init__.py,sha256=uFQW_cJyymxtHQiYb_xjzZAklc487L0n9GkgHgC7yAI,2618
 ingestr/src/freshdesk/freshdesk_client.py,sha256=3z5Yc008ADzRcJWtNc00PwjkLzG-RMI8jVIOOyYA-Rw,4088
 ingestr/src/freshdesk/settings.py,sha256=0Wr_OMnUZcTlry7BmALssLxD2yh686JW4moLNv12Jnw,409
@@ -87,6 +87,8 @@ ingestr/src/notion/helpers/client.py,sha256=QXuudkf5Zzff98HRsCqA1g1EZWIrnfn1falP
 ingestr/src/notion/helpers/database.py,sha256=gigPibTeVefP3lA-8w4aOwX67pj7RlciPk5koDs1ry8,2737
 ingestr/src/personio/__init__.py,sha256=sHYpoV-rg-kA1YsflctChis0hKcTrL6mka9O0CHV4zA,11638
 ingestr/src/personio/helpers.py,sha256=EKmBN0Lf4R0lc3yqqs7D-RjoZ75E8gPcctt59xwHxrY,2901
+ingestr/src/phantombuster/__init__.py,sha256=5XGwbtNYmRxL4lmcC0TmyXgljY_21DvDIdw9zaWmgvI,1757
+ingestr/src/phantombuster/client.py,sha256=WO87AGU3Fphd9DgWTkrSAfldLH1XxcPxvNVmNxQ_5eU,2785
 ingestr/src/pipedrive/__init__.py,sha256=iRrxeMwo8_83ptgGnTFTNHV1nYvIsFfg0a3XzugPYeI,6982
 ingestr/src/pipedrive/settings.py,sha256=q119Fy4C5Ip1rMoCILX2BkHV3bwiXC_dW58KIiDUzsY,708
 ingestr/src/pipedrive/typing.py,sha256=lEMXu4hhAA3XkhVSlBUa-juqyupisd3c-qSQKxFvzoE,69
@@ -125,8 +127,8 @@ ingestr/testdata/delete_insert_part2.csv,sha256=B_KUzpzbNdDY_n7wWop1mT2cz36TmayS
 ingestr/testdata/merge_expected.csv,sha256=DReHqWGnQMsf2PBv_Q2pfjsgvikYFnf1zYcQZ7ZqYN0,276
 ingestr/testdata/merge_part1.csv,sha256=Pw8Z9IDKcNU0qQHx1z6BUf4rF_-SxKGFOvymCt4OY9I,185
 ingestr/testdata/merge_part2.csv,sha256=T_GiWxA81SN63_tMOIuemcvboEFeAmbKc7xRXvL9esw,287
-ingestr-0.13.36.dist-info/METADATA,sha256=AFJ4qtGMrtaG5luUcRCXAsp7yP8FKlL4EjP8KorvXKI,13575
-ingestr-0.13.36.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-ingestr-0.13.36.dist-info/entry_points.txt,sha256=oPJy0KBnPWYjDtP1k8qwAihcTLHSZokSQvRAw_wtfJM,46
-ingestr-0.13.36.dist-info/licenses/LICENSE.md,sha256=cW8wIhn8HFE-KLStDF9jHQ1O_ARWP3kTpk_-eOccL24,1075
-ingestr-0.13.36.dist-info/RECORD,,
+ingestr-0.13.38.dist-info/METADATA,sha256=ZknSXuae-_lb60AWDACr731hTr0Q_bX_9sUds6rHM28,13575
+ingestr-0.13.38.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+ingestr-0.13.38.dist-info/entry_points.txt,sha256=oPJy0KBnPWYjDtP1k8qwAihcTLHSZokSQvRAw_wtfJM,46
+ingestr-0.13.38.dist-info/licenses/LICENSE.md,sha256=cW8wIhn8HFE-KLStDF9jHQ1O_ARWP3kTpk_-eOccL24,1075
+ingestr-0.13.38.dist-info/RECORD,,