ingestr 0.14.94__py3-none-any.whl → 0.14.96__py3-none-any.whl

ingestr/src/buildinfo.py

@@ -1 +1 @@
-version = "v0.14.94"
+version = "v0.14.96"

ingestr/src/factory.py

@@ -84,6 +84,7 @@ from ingestr.src.sources import (
     ShopifySource,
     SlackSource,
     SmartsheetSource,
+    SocrataSource,
     SolidgateSource,
     SqlSource,
     StripeAnalyticsSource,
@@ -218,6 +219,7 @@ class SourceDestinationFactory:
         "sftp": SFTPSource,
         "pinterest": PinterestSource,
         "revenuecat": RevenueCatSource,
+        "socrata": SocrataSource,
         "zoom": ZoomSource,
         "clickup": ClickupSource,
         "influxdb": InfluxDBSource,

ingestr/src/socrata_source/__init__.py

@@ -0,0 +1,83 @@
+"""A source loading data from Socrata open data platform"""
+
+from typing import Any, Dict, Iterator, Optional
+
+import dlt
+
+from .helpers import fetch_data
+
+
+@dlt.source(name="socrata", max_table_nesting=0)
+def source(
+    domain: str,
+    dataset_id: str,
+    app_token: Optional[str] = None,
+    username: Optional[str] = None,
+    password: Optional[str] = None,
+    incremental: Optional[Any] = None,
+    primary_key: Optional[str] = None,
+    write_disposition: Optional[str] = dlt.config.value,
+):
+    """
+    A dlt source for the Socrata open data platform.
+
+    Supports both full refresh (replace) and incremental loading (merge).
+
+    Args:
+        domain: The Socrata domain (e.g., "evergreen.data.socrata.com")
+        dataset_id: The dataset identifier (e.g., "6udu-fhnu")
+        app_token: Socrata app token for higher rate limits (recommended)
+        username: Username for authentication (if dataset is private)
+        password: Password for authentication (if dataset is private)
+        incremental: DLT incremental object for incremental loading
+        primary_key: Primary key field for merge operations (default: ":id")
+        write_disposition: Write disposition ("replace", "append", "merge").
+            If not provided, automatically determined based on incremental setting.
+
+    Returns:
+        A dlt source with a single "dataset" resource
+    """
+
+    @dlt.resource(
+        write_disposition=write_disposition or "replace",
+        primary_key=primary_key,  # type: ignore[call-overload]
+    )
+    def dataset(
+        incremental: Optional[dlt.sources.incremental] = incremental,  # type: ignore[type-arg]
+    ) -> Iterator[Dict[str, Any]]:
+        """
+        Yields records from a Socrata dataset.
+
+        Supports both full refresh (replace) and incremental loading (merge).
+        When incremental is provided, filters data using SoQL WHERE clause on the server side.
+
+        Yields:
+            Dict[str, Any]: Individual records from the dataset
+        """
+        fetch_kwargs: Dict[str, Any] = {
+            "domain": domain,
+            "dataset_id": dataset_id,
+            "app_token": app_token,
+            "username": username,
+            "password": password,
+        }
+
+        if incremental and incremental.cursor_path:
+            fetch_kwargs["incremental_key"] = incremental.cursor_path
+            fetch_kwargs["start_value"] = (
+                str(incremental.last_value)
+                if incremental.last_value is not None
+                else None
+            )
+            if getattr(incremental, "end_value", None) is not None:
+                ev = incremental.end_value  # type: ignore[attr-defined]
+                fetch_kwargs["end_value"] = (
+                    ev.isoformat()  # type: ignore[union-attr]
+                    if hasattr(ev, "isoformat")
+                    else str(ev)
+                )
+
+        # Fetch and yield records
+        yield from fetch_data(**fetch_kwargs)
+
+    return (dataset,)

ingestr/src/socrata_source/helpers.py

@@ -0,0 +1,85 @@
+"""Socrata API helpers"""
+
+from typing import Any, Dict, Iterator, Optional
+
+from dlt.sources.helpers import requests
+
+from .settings import DEFAULT_PAGE_SIZE, REQUEST_TIMEOUT
+
+
+def fetch_data(
+    domain: str,
+    dataset_id: str,
+    app_token: Optional[str] = None,
+    username: Optional[str] = None,
+    password: Optional[str] = None,
+    incremental_key: Optional[str] = None,
+    start_value: Optional[str] = None,
+    end_value: Optional[str] = None,
+) -> Iterator[Dict[str, Any]]:
+    """
+    Fetch records from Socrata dataset with pagination and optional filtering.
+
+    Uses offset-based pagination to get all records, not just first 50000.
+    Supports incremental loading via SoQL WHERE clause for server-side filtering.
+
+    Args:
+        domain: Socrata domain (e.g., "data.seattle.gov")
+        dataset_id: Dataset identifier (e.g., "6udu-fhnu")
+        app_token: Socrata app token for higher rate limits
+        username: Username for authentication
+        password: Password for authentication
+        start_value: Minimum value for incremental_key (inclusive)
+        end_value: Maximum value for incremental_key (exclusive)
+
+    Yields:
+        Lists of records (one list per page)
+
+    Raises:
+        requests.HTTPError: If API request fails
+    """
+    url = f"https://{domain}/resource/{dataset_id}.json"
+
+    headers = {"Accept": "application/json"}
+    if app_token:
+        headers["X-App-Token"] = app_token
+
+    auth = (username, password) if username and password else None
+
+    limit = DEFAULT_PAGE_SIZE
+    offset = 0
+
+    while True:
+        params: Dict[str, Any] = {"$limit": limit, "$offset": offset}
+
+        if incremental_key and start_value:
+            start_value_iso = str(start_value).replace(" ", "T")
+            where_conditions = [f"{incremental_key} >= '{start_value_iso}'"]
+
+            if end_value:
+                end_value_iso = str(end_value).replace(" ", "T")
+                where_conditions.append(f"{incremental_key} < '{end_value_iso}'")
+
+            params["$where"] = " AND ".join(where_conditions)
+            params["$order"] = f"{incremental_key} ASC"
+
+        response = requests.get(
+            url,
+            headers=headers,
+            auth=auth,
+            params=params,
+            timeout=REQUEST_TIMEOUT,
+        )
+        response.raise_for_status()
+
+        data = response.json()
+
+        if not data:
+            break
+
+        yield data
+
+        if len(data) < limit:
+            break
+
+        offset += limit

ingestr/src/socrata_source/settings.py

@@ -0,0 +1,8 @@
+"""Socrata API settings and constants"""
+
+# Request timeout in seconds
+REQUEST_TIMEOUT = 30
+
+# Maximum number of records to fetch per page
+# Socrata API supports up to 50000 records per request
+DEFAULT_PAGE_SIZE = 50000

ingestr/src/sources.py

@@ -4233,3 +4233,93 @@ class CouchbaseSource:
         table_instance.max_table_nesting = 1
 
         return table_instance
+
+
+class SocrataSource:
+    def handles_incrementality(self) -> bool:
+        return False
+
+    def dlt_source(self, uri: str, table: str, **kwargs):
+        """
+        Creates a DLT source for Socrata open data platform.
+
+        URI format: socrata://domain?app_token=TOKEN
+        Table: dataset_id (e.g., "6udu-fhnu")
+
+        Args:
+            uri: Socrata connection URI with domain and optional auth params
+            table: Dataset ID (e.g., "6udu-fhnu")
+            **kwargs: Additional arguments:
+                - incremental_key: Field to use for incremental loading (e.g., ":updated_at")
+                - interval_start: Start date for initial load
+                - interval_end: End date for load
+                - primary_key: Primary key field for merge operations
+
+        Returns:
+            DltResource for the Socrata dataset
+        """
+        from urllib.parse import parse_qs, urlparse
+
+        parsed = urlparse(uri)
+
+        domain = parsed.netloc
+        if not domain:
+            raise ValueError(
+                "Domain must be provided in the URI.\n"
+                "Format: socrata://domain?app_token=TOKEN\n"
+                "Example: socrata://evergreen.data.socrata.com?app_token=mytoken"
+            )
+
+        query_params = parse_qs(parsed.query)
+
+        dataset_id = table
+        if not dataset_id:
+            raise ValueError(
+                "Dataset ID must be provided as the table parameter.\n"
+                "Example: --source-table 6udu-fhnu"
+            )
+
+        app_token = query_params.get("app_token", [None])[0]
+        username = query_params.get("username", [None])[0]
+        password = query_params.get("password", [None])[0]
+
+        incremental = None
+        if kwargs.get("incremental_key"):
+            start_value = kwargs.get("interval_start")
+            end_value = kwargs.get("interval_end")
+
+            if start_value:
+                start_value = (
+                    start_value.isoformat()
+                    if hasattr(start_value, "isoformat")
+                    else str(start_value)
+                )
+
+            if end_value:
+                end_value = (
+                    end_value.isoformat()
+                    if hasattr(end_value, "isoformat")
+                    else str(end_value)
+                )
+
+            incremental = dlt_incremental(
+                kwargs.get("incremental_key", ""),
+                initial_value=start_value,
+                end_value=end_value,
+                range_end="open",
+                range_start="closed",
+            )
+
+        primary_key = kwargs.get("primary_key")
+
+        from ingestr.src.socrata_source import source
+
+        return source(
+            domain=domain,
+            dataset_id=dataset_id,
+            app_token=app_token,
+            username=username,
+            password=password,
+            incremental=incremental,
+            primary_key=primary_key,
+        ).with_resources("dataset")

{ingestr-0.14.94.dist-info → ingestr-0.14.96.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ingestr
-Version: 0.14.94
+Version: 0.14.96
 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-0.14.94.dist-info → ingestr-0.14.96.dist-info}/RECORD

@@ -2,17 +2,17 @@ ingestr/conftest.py,sha256=OE2yxeTCosS9CUFVuqNypm-2ftYvVBeeq7egm3878cI,1981
 ingestr/main.py,sha256=qo0g3wCFl8a_1jUwXagX8L1Q8PKKQlTF7md9pfnzW0Y,27155
 ingestr/src/.gitignore,sha256=8cX1AZTSI0TcdZFGTmS_oyBjpfCzhOEt0DdAo2dFIY8,203
 ingestr/src/blob.py,sha256=UUWMjHUuoR9xP1XZQ6UANQmnMVyDx3d0X4-2FQC271I,2138
-ingestr/src/buildinfo.py,sha256=CcaXu7ayEOiHaFuTFCgGdFP5zqt7HZWvn7bvO_OyRhk,21
+ingestr/src/buildinfo.py,sha256=-9qPR_WQg9aaTRg324DJAZs43V_FQHsRu9G9xDfXrjE,21
 ingestr/src/destinations.py,sha256=QtjE0AGs0WkPHaI2snWPHJ8HHi4lwXUBYLJPklz8Mvk,27772
 ingestr/src/errors.py,sha256=fhJ2BxOqOsBfOxuTDKfZblvawBrPG3x_1VikIxMZBRI,874
-ingestr/src/factory.py,sha256=WBc2y5N-9HH3WOTUSe7sDEftarScDAipka7CSW-A4L4,7829
+ingestr/src/factory.py,sha256=iFOFbwifvQf7qOtSoNPS6RGvAhsRaX7HzbjouHmSvfs,7882
 ingestr/src/filters.py,sha256=0n0sNAVG_f-B_1r7lW5iNtw9z_G1bxWzPaiL1i6tnbU,1665
 ingestr/src/http_client.py,sha256=bxqsk6nJNXCo-79gW04B53DQO-yr25vaSsqP0AKtjx4,732
 ingestr/src/loader.py,sha256=9NaWAyfkXdqAZSS-N72Iwo36Lbx4PyqIfaaH1dNdkFs,1712
 ingestr/src/masking.py,sha256=VN0LdfvExhQ1bZMRylGtaBUIoH-vjuIUmRnYKwo3yiY,11358
 ingestr/src/partition.py,sha256=BrIP6wFJvyR7Nus_3ElnfxknUXeCipK_E_bB8kZowfc,969
 ingestr/src/resource.py,sha256=ZqmZxFQVGlF8rFPhBiUB08HES0yoTj8sZ--jKfaaVps,1164
-ingestr/src/sources.py,sha256=XEimy9ph8QDW8vjNocjovxM6cdrUkIPmFGZ9eLXZWVk,148842
+ingestr/src/sources.py,sha256=JVZf22XgIFXov3-yKOjsbQVw9cV_LrDeXD6eb4Z6jFk,151802
 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
@@ -154,6 +154,9 @@ 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=RIEfN1T2TMFg8T0RvN4o6sqC58YusJRDrmE9Isos5P4,2375
+ingestr/src/socrata_source/__init__.py,sha256=K5DVpsVXTMfunZd5YoEsn1nipfo1zavFS59g3m2tsc8,2984
+ingestr/src/socrata_source/helpers.py,sha256=KbVojFSmMLXb0ajh8bhqfZfxDHH7rQ3nyI8p2jxVifA,2500
+ingestr/src/socrata_source/settings.py,sha256=DLfu-4HOa5nR7h9tbOySEa2ye3w_Z6TYZ9_zPqWaNQk,220
 ingestr/src/solidgate/__init__.py,sha256=Ts83j-JSnFsFuF4tDhVOfZKg7H0-bIpfn3kg1ZOR58A,8003
 ingestr/src/solidgate/helpers.py,sha256=mAsW_1hpD7ab3Y2vw8fxHi4yD3aT1geLdIYZ7ycyxBc,5690
 ingestr/src/sql_database/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -186,8 +189,8 @@ 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=zf3DXT29Y4TH2lNPBFphdjlaelUUyPJcsW2UO68RzDs,4862
-ingestr-0.14.94.dist-info/METADATA,sha256=B9NePP8xKNVs7YZ6hZV6gDfTGBxdRwxs7aMd5xPL6As,15359
-ingestr-0.14.94.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-ingestr-0.14.94.dist-info/entry_points.txt,sha256=oPJy0KBnPWYjDtP1k8qwAihcTLHSZokSQvRAw_wtfJM,46
-ingestr-0.14.94.dist-info/licenses/LICENSE.md,sha256=cW8wIhn8HFE-KLStDF9jHQ1O_ARWP3kTpk_-eOccL24,1075
-ingestr-0.14.94.dist-info/RECORD,,
+ingestr-0.14.96.dist-info/METADATA,sha256=vnkdaQVPvlnpHq9UgecuzRSSb_IiKE6_gS1jLkYzGEY,15359
+ingestr-0.14.96.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+ingestr-0.14.96.dist-info/entry_points.txt,sha256=oPJy0KBnPWYjDtP1k8qwAihcTLHSZokSQvRAw_wtfJM,46
+ingestr-0.14.96.dist-info/licenses/LICENSE.md,sha256=cW8wIhn8HFE-KLStDF9jHQ1O_ARWP3kTpk_-eOccL24,1075
+ingestr-0.14.96.dist-info/RECORD,,