feat: S3 batch storage

cookiecutter/tap-template/{{cookiecutter.tap_id}}/pyproject.toml

@@ -12,7 +12,8 @@ license = "Apache 2.0"
 [tool.poetry.dependencies]
 python = "<3.11,>=3.7.1"
 requests = "^2.25.1"
-singer-sdk = "^0.14.0"
+singer-sdk = { version="^0.14.0"}
+fs-s3fs = { version = "^1.1.1", optional = true}
 
 [tool.poetry.dev-dependencies]
 pytest = "^6.2.5"
@@ -24,6 +25,9 @@ mypy = "^0.910"
 types-requests = "^2.26.1"
 isort = "^5.10.1"
 
+[tool.poetry.extras]
+s3 = ["fs-s3fs"]
+
 [tool.isort]
 profile = "black"
 multi_line_output = 3 # Vertical Hanging Indent

cookiecutter/target-template/{{cookiecutter.target_id}}/pyproject.toml

@@ -12,7 +12,8 @@ license = "Apache 2.0"
 [tool.poetry.dependencies]
 python = "<3.11,>=3.7.1"
 requests = "^2.25.1"
-singer-sdk = "^0.14.0"
+singer-sdk = { version="^0.14.0"}
+fs-s3fs = { version = "^1.1.1", optional = true}
 
 [tool.poetry.dev-dependencies]
 pytest = "^6.2.5"
@@ -24,6 +25,9 @@ mypy = "^0.910"
 types-requests = "^2.26.1"
 isort = "^5.10.1"
 
+[tool.poetry.extras]
+s3 = ["fs-s3fs"]
+
 [tool.isort]
 profile = "black"
 multi_line_output = 3 # Vertical Hanging Indent

docs/batch.md

@@ -14,10 +14,11 @@ This library's implementation of the `BATCH` message is used to send records in
 - when the tap outputs records at a much higher rate than the target can consume them, creating backpressure
 - when the source system can directly export data in bulk (e.g. a database dump)
 
-Currently only a local filesystem is supported, but other filesystems like AWS S3, FTP, etc. could be supported in the future.
+Currently only a local filesystem or AWS S3 are supported, but other filesystems like FTP, etc. could be supported in the future.
 
 ## The `BATCH` Message
 
+Local
 ```json
 {
   "type": "BATCH",
@@ -33,6 +34,22 @@ Currently only a local filesystem is supported, but other filesystems like AWS S
 }
 ```
 
+AWS S3
+```json
+{
+  "type": "BATCH",
+  "stream": "users",
+  "encoding": {
+    "format": "jsonl",
+    "compression": "gzip"
+  },
+  "manifest": [
+    "s3://path/to/batch/file/1",
+    "s3://path/to/batch/file/2"
+  ]
+}
+```
+
 ### `encoding`
 
 The `encoding` field is used to specify the format and compression of the batch files. Currently only `jsonl` and `gzip` are supported, respectively.
@@ -43,7 +60,7 @@ The `manifest` field is used to specify the paths to the batch files. The paths
 
 ## Batch configuration
 
-When local storage is used, targets do no require special configuration to process `BATCH` messages.
+When local storage is used, targets do no require special configuration to process `BATCH` messages. Use of AWS S3 assumes S3/AWS credentials are already discoverable via the underlying S3 libraries (`AWS_ACCESS_KEY_ID` / `AWS_SECRET_ACCESS_KEY` / `AWS_DEFAULT_REGION`)
 
 Taps may be configured to specify a root storage `root` directory, file path `prefix`, and `encoding` for batch files using a configuration like the below:
 

noxfile.py

@@ -69,7 +69,8 @@ def mypy(session: Session) -> None:
 @session(python=python_versions)
 def tests(session: Session) -> None:
     """Execute pytest tests and compute coverage."""
-    session.install(".")
+
+    session.install(".[s3]")
     session.install(*test_dependencies)
 
     # temp fix until pyarrow is supported on python 3.11

pyproject.toml

@@ -68,6 +68,9 @@ sphinx-copybutton = {version = ">=0.3.1,<0.6.0", optional = true}
 myst-parser = {version = ">=0.17.2,<0.19.0", optional = true}
 sphinx-autobuild = {version = "^2021.3.14", optional = true}
 
+# File storage dependencies installed as optional 'filesystem' extras
+fs-s3fs = { version = "^1.1.1", optional = true}
+
 [tool.poetry.extras]
 docs = [
     "sphinx",
@@ -76,6 +79,7 @@ docs = [
     "myst-parser",
     "sphinx-autobuild",
 ]
+s3 = ["fs-s3fs"]
 # Mark these as extras so that installs pass on python 3.11
 samples = [
     "pyarrow",

singer_sdk/helpers/_resources.py

@@ -2,6 +2,7 @@
 
 import sys
 from types import ModuleType
+from typing import cast
 
 if sys.version_info >= (3, 9):
     import importlib.resources as importlib_resources
@@ -21,4 +22,4 @@ def get_package_files(package: str | ModuleType) -> Traversable:
     Returns:
         The file as a Traversable object.
     """
-    return importlib_resources.files(package)
+    return cast(Traversable, importlib_resources.files(package))

tests/core/test_batch.py

@@ -32,6 +32,16 @@ def test_storage_get_url():
         assert url.replace("\\", "/").endswith("root_dir/prefix--file.jsonl.gz")
 
 
+def test_storage_get_s3_url():
+    storage = StorageTarget("s3://testing123:testing123@test_bucket")
+
+    with storage.fs(create=True) as fs:
+        url = fs.geturl("prefix--file.jsonl.gz")
+        assert url.startswith(
+            "https://s3.amazonaws.com/test_bucket/prefix--file.jsonl.gz"
+        )
+
+
 @pytest.mark.parametrize(
     "file_url,root",
     [
@@ -40,6 +50,11 @@ def test_storage_get_url():
             "file:///Users/sdk/path/to",
             id="local",
         ),
+        pytest.param(
+            "s3://test_bucket/prefix--file.jsonl.gz",
+            "s3://test_bucket",
+            id="s3",
+        ),
     ],
 )
 def test_storage_from_url(file_url: str, root: str):