Validate format of tags to be max 255 characters and with certain characters (#23)

Match up tag validations that came in with the main project here [1].
The multi insertion queries use commas to split tags during batch
inserts, so it's important that incoming tags don't have comms of their
own.

[1] riverqueue/river#351

Author: brandur

CHANGELOG.md

@@ -7,29 +7,33 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Changed
+
+- Tags are now limited to 255 characters in length, and should match the regex `\A[\w][\w\-]+[\w]\z` (importantly, they can't contain commas). [PR #23](https://github.com/riverqueue/riverqueue-python/pull/23).
+
 ## [0.3.0] - 2024-07-04
 
 ### Added
 
-- Implement `insert_many` and `insert_many_tx`. [PR #22](https://github.com/riverqueue/river/pull/22).
+- Implement `insert_many` and `insert_many_tx`. [PR #22](https://github.com/riverqueue/riverqueue-python/pull/22).
 
 ## [0.2.0] - 2024-07-04
 
 ### Changed
 
-- Rename `Args` to `JobArgs` and add `JobArgsWithInsertOpts` protocol. [PR #20](https://github.com/riverqueue/river/pull/20).
+- Rename `Args` to `JobArgs` and add `JobArgsWithInsertOpts` protocol. [PR #20](https://github.com/riverqueue/riverqueue-python/pull/20).
 
 ## [0.1.2] - 2024-07-04
 
 ### Changed
 
-- Add usage instructions README, add job state constants, and change return value of `insert_many()` and `insert_many_tx()` to an integer instead of a list of jobs. [PR #19](https://github.com/riverqueue/river/pull/19).
+- Add usage instructions README, add job state constants, and change return value of `insert_many()` and `insert_many_tx()` to an integer instead of a list of jobs. [PR #19](https://github.com/riverqueue/riverqueue-python/pull/19).
 
 ## [0.1.1] - 2024-07-04
 
 ### Fixed
 
-- Fix `pyproject.toml` description and add various URLs like to homepage, docs, and GitHub repositories. [PR #18](https://github.com/riverqueue/river/pull/18).
+- Fix `pyproject.toml` description and add various URLs like to homepage, docs, and GitHub repositories. [PR #18](https://github.com/riverqueue/riverqueue-python/pull/18).
 
 ## [0.1.0] - 2024-07-04
 

src/riverqueue/client.py

@@ -1,5 +1,6 @@
 from dataclasses import dataclass
 from datetime import datetime, timezone, timedelta
+import re
 from typing import (
     Any,
     Awaitable,
@@ -320,7 +321,7 @@ def _make_insert_params(
         queue=insert_opts.queue or args_insert_opts.queue or QUEUE_DEFAULT,
         scheduled_at=scheduled_at and scheduled_at.astimezone(timezone.utc),
         state="scheduled" if scheduled_at else "available",
-        tags=insert_opts.tags or args_insert_opts.tags or [],
+        tags=_validate_tags(insert_opts.tags or args_insert_opts.tags or []),
     )
 
     return insert_params, unique_opts
@@ -348,3 +349,14 @@ def _truncate_time(time, interval_seconds) -> datetime:
 def _uint64_to_int64(uint64):
     # Packs a uint64 then unpacks to int64 to fit within Postgres bigint
     return (uint64 + (1 << 63)) % (1 << 64) - (1 << 63)
+
+
+tag_re = re.compile("\A[\w][\w\-]+[\w]\Z")
+
+
+def _validate_tags(tags: list[str]) -> list[str]:
+    for tag in tags:
+        assert (
+            len(tag) <= 255 and tag_re.match(tag)
+        ), f"tags should be less than 255 characters in length and match regex {tag_re.pattern}"
+    return tags

tests/client_test.py

@@ -259,6 +259,26 @@ def to_json() -> None:
     assert "args should return non-nil from `to_json`" == str(ex.value)
 
 
+def test_tag_validation(client):
+    client.insert(
+        SimpleArgs(), insert_opts=InsertOpts(tags=["foo", "bar", "baz", "foo-bar-baz"])
+    )
+
+    with pytest.raises(AssertionError) as ex:
+        client.insert(SimpleArgs(), insert_opts=InsertOpts(tags=["commas,bad"]))
+    assert (
+        "tags should be less than 255 characters in length and match regex \A[\w][\w\-]+[\w]\Z"
+        == str(ex.value)
+    )
+
+    with pytest.raises(AssertionError) as ex:
+        client.insert(SimpleArgs(), insert_opts=InsertOpts(tags=["a" * 256]))
+    assert (
+        "tags should be less than 255 characters in length and match regex \A[\w][\w\-]+[\w]\Z"
+        == str(ex.value)
+    )
+
+
 def test_check_advisory_lock_prefix_bounds():
     Client(mock_driver, advisory_lock_prefix=123)