ingestr 0.13.36__tar.gz → 0.13.38__tar.gz

{ingestr-0.13.36 → ingestr-0.13.38}/PKG-INFO

@@ -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 → ingestr-0.13.38}/docs/.vitepress/config.mjs

@@ -119,6 +119,7 @@ export default defineConfig({
                 link: "/supported-sources/facebook-ads.md",
               },
               { text: "Frankfurter", link: "/supported-sources/frankfurter.md" },
+              { text: "Freshdesk", link: "/supported-sources/freshdesk.md" },
               { text: "Google Cloud Storage (GCS)", link: "/supported-sources/gcs.md" },
               { text: "Google Analytics", link: "/supported-sources/google_analytics.md" },
               { text: "Google Ads", link: "/supported-sources/google-ads.md" },
@@ -130,6 +131,7 @@ export default defineConfig({
               { text: "LinkedIn Ads", link: "/supported-sources/linkedin_ads.md" },
               { text: "Notion", link: "/supported-sources/notion.md" },
               { text: "Personio", link: "/supported-sources/personio.md" },
+              { text: "PhantomBuster", link: "/supported-sources/phantombuster.md" },
               { text: "Pipedrive", link: "/supported-sources/pipedrive.md" },
               { text: "S3", link: "/supported-sources/s3.md" },
               { text: "Salesforce", link: "/supported-sources/salesforce.md" },

{ingestr-0.13.36 → ingestr-0.13.38}/docs/supported-sources/frankfurter.md

@@ -10,7 +10,7 @@ The `ingestr` command to use the `frankfurter` source is as follows:
 
 ```bash
 ingestr ingest \
---source-uri 'frankfurter://' \
+--source-uri 'frankfurter://?base=IDR' \
 --interval-start '2025-03-27' \     # Optional. See 'exchange_rates'.
 --interval-end '2025-03-28' \       # Optional.
 --source-table '<table_name>' \     # E.g 'currencies', 'latest', 'exchange_rates'. See below.
@@ -25,7 +25,9 @@ ingestr ingest \
 ### **`--source-uri`**
 - **Description**: Specifies the source URI for the Frankfurter API.
 - **Value**: `'frankfurter://'`
-- **Purpose**: Indicates that the data will be fetched from the Frankfurter API.
+- **Purpose**: Indicates that the data will be fetched from the Frankfurter API. 
+  - An optional base currency can be added `?base={base_currency}`.
+  - If no base currency is included, base currency defaults to USD.
 
 ---
 
@@ -89,9 +91,10 @@ ingestr ingest \
 - **Description**: Fetches the latest exchange rates.
 - **Columns**:
   - `date`: The date of the exchange rates.
-  - `currency_name`: The ISO 4217 currency code (e.g., `USD`, `EUR`).
+  - `currency_code`: The ISO 4217 currency code (e.g., `USD`, `EUR`).
   - `rate`: The exchange rate relative to the base currency.
-- **Primary Key**: Composite key of `date` and `currency_name`.
+  - `base_currency`: The base currency used to calculate the exchange rate.
+- **Primary Key**: Composite key of `date`, `currency_code` and `base_currency`.
 - **Notes**:
   - The base currency (e.g., `EUR`) is included with a rate of `1.0`.
 
@@ -101,9 +104,10 @@ ingestr ingest \
 - **Description**: Fetches historical exchange rates for a specified date range.
 - **Columns**:
   - `date`: The date of the exchange rates.
-  - `currency_name`: The ISO 4217 currency code (e.g., `USD`, `EUR`).
+  - `currency_code`: The ISO 4217 currency code (e.g., `USD`, `EUR`).
   - `rate`: The exchange rate relative to the base currency.
-- **Primary Key**: Composite key of `date` and `currency_name`.
+  - `base_currency`: The base currency used to calculate the exchange rate.
+- **Primary Key**: Composite key of `date`, `currency_code` and `base_currency`.
 - **Notes**:
   - An optional start and end date can be added via the arguments `--interval-start` and optionally `--interval-end` to define the date range (see examples below). If no start date is specified, the date will default today's date (and thus return the latest exchange rates).
   - If a start date but no end date is specified, then the end date will default to today's date and ingestr will retrieve data up until the latest published data.
@@ -114,24 +118,24 @@ Here `--interval-start` is set to a weekend date (e.g., `2025-03-29` -- a Saturd
 
 `--interval-start` defaults to the previous Friday (`2025-03-28`) and the next data is from the following Monday (for simplicity, only a subset of currencies is shown below):
 
-| **date**     | **currency_name** | **rate** |
-|--------------|-------------------|----------|
-| 2025-03-28   | EUR               | 1.0      | 
-| 2025-03-28   | USD               | 1.0783   | 
-| 2025-03-28   | GBP               | 0.8571   | 
-| 2025-03-31   | EUR               | 1.0      | 
-| 2025-03-31   | USD               | 1.0783   |
-| 2025-03-31   | GBP               | 0.8571   |
+| **date**     | **currency_code** | **rate** | **base_currency** |
+|--------------|-------------------|----------|-------------------|
+| 2025-03-28   | EUR               | 1.0      | EUR               |
+| 2025-03-28   | USD               | 1.0783   | EUR               |
+| 2025-03-28   | GBP               | 0.8571   | EUR               |
+| 2025-03-31   | EUR               | 1.0      | EUR               |
+| 2025-03-31   | USD               | 1.0783   | EUR               |
+| 2025-03-31   | GBP               | 0.8571   | EUR               |
 
 
 ---
 
 ## **Examples**
 
-### **1. Fetch the Latest Exchange Rates**
+### **1. Fetch the Latest Exchange Rates with GBP as Base Currency**
 ```bash
 ingestr ingest \
---source-uri 'frankfurter://' \
+--source-uri 'frankfurter://?base=GBP' \
 --source-table 'latest' \
 --dest-uri 'duckdb.db' \
 --dest-table 'schema.latest_new_scheme'
@@ -139,7 +143,7 @@ ingestr ingest \
 
 ---
 
-### **2. Fetch Historical Exchange Rates**
+### **2. Fetch Historical Exchange Rates with USD as Default Base Currency**
 ```bash
 ingestr ingest \
 --source-uri 'frankfurter://' \

ingestr-0.13.38/docs/supported-sources/phantombuster.md

@@ -0,0 +1,38 @@
+# PhantomBuster
+[PhantomBuster](https://phantombuster.com/) is a cloud-based data automation and web scraping platform that allows users to extract data from websites, automate actions.
+
+ingestr supports PhantomBuster as a source.
+
+## URI format
+
+The URI format for PhantomBuster is as follows:
+
+```plaintext
+PhantomBuster://?api_key=<api_key>
+```
+
+URI parameters:
+- `api_key`: the API key used for authentication with the PhantomBuster API
+
+## Setting up a PhantomBuster Integration
+
+You can find your PhantomBuster API key by following the guide [here](https://hub.phantombuster.com/docs/api#how-to-find-my-api-key).
+
+Let's say your `api_key` is key_123, here's a sample command that will copy the data from PhantomBuster into a DuckDB database:
+
+
+```bash
+ingestr ingest \
+--source-uri 'PhantomBuster://?api_key=key_123' \
+--source-table 'completed_phantoms:<agent_id>' \
+--dest-uri duckdb:///PhantomBuster.duckdb \
+--dest-table 'dest.result'
+```
+
+
+<img alt="PhantomBuster_img" src="../media/phantombuster.png"/>
+
+
+For now, we only support `completed_phantoms` table followed by an `agent_id`. For example: `completed_phantoms:<agent_id>` Where agent id is a unique identifier for a specific Phantom which can be found in URI of a specific phantom.
+
+Use this as `--source-table` parameter in the `ingestr ingest` command.

{ingestr-0.13.36 → ingestr-0.13.38}/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-0.13.38/ingestr/src/buildinfo.py

@@ -0,0 +1 @@
+version = "v0.13.38"

{ingestr-0.13.36 → ingestr-0.13.38}/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-0.13.36 → ingestr-0.13.38}/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-0.13.36 → ingestr-0.13.38}/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-0.13.38/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-0.13.38/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-0.13.36 → ingestr-0.13.38}/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 → ingestr-0.13.38}/requirements.txt

@@ -187,7 +187,7 @@ googleapis-common-protos==1.69.0
     #   google-ads
     #   google-api-core
     #   grpcio-status
-greenlet==3.2.1
+greenlet==3.2.2
     # via sqlalchemy
 grpcio==1.70.0
     # via

ingestr-0.13.36/ingestr/src/buildinfo.py

@@ -1 +0,0 @@
-version = "v0.13.36"