Harden Rhiza sync schedule editing and tighten exception API typing

[Copilot]

This PR applies the requested fixes across Rhiza make targets and exception typing. It removes fragile in-place sed backup behavior, corrects command chaining in summarise-sync, and refines exception signatures/payload handling in exceptions.py.

• Rhiza sync workflow editing

  • Replaced sed -i.bak ... && rm ...bak with a temp-file rewrite + atomic move pattern in _apply-sync-schedule.
  • Avoids backup-file residue and removes BSD/GNU sed -i portability concerns.

• Rhiza summarise target command flow

  • Updated summarise-sync to use $(MAKE) install-uv && ... instead of ;.
  • Ensures the summarise command only runs when install-uv succeeds.

• Exception API/type adjustments

  • InvalidMaxBpsError.__init__ now uses Any for clearer intent than object.
  • MissingDateColumnError now consistently stores available as a list via defensive normalization/copying.

• Focused exception coverage

  • Added targeted tests for MissingDateColumnError.available default/list-copy behavior.

# before
sed -i.bak "s|cron: '[^']*'|cron: '$(RHIZA_SYNC_SCHEDULE)'|" .github/workflows/rhiza_sync.yml && rm -f .github/workflows/rhiza_sync.yml.bak

# after
sed "s|cron: '[^']*'|cron: '$(RHIZA_SYNC_SCHEDULE)'|" .github/workflows/rhiza_sync.yml > .github/workflows/rhiza_sync.yml.tmp && \
mv .github/workflows/rhiza_sync.yml.tmp .github/workflows/rhiza_sync.yml

Original prompt

Please apply the following diffs and create a pull request.
Once the PR is ready, give it a title based on the messages of the fixes being applied.

[{"message":"The `sed -i.bak` invocation creates a backup file and then removes it. On macOS (BSD sed), `-i ''` is required for in-place editing without a backup, while GNU sed uses `-i` alone. Using `-i.bak` and then deleting the backup is a common cross-platform workaround, but if `sed` fails mid-run the `.bak` file may be left behind. A more robust approach would be to check whether the substitution actually changed the file (e.g. via `diff`) or to use a temp file pattern.","fixFiles":[{"filePath":".rhiza/rhiza.mk","diff":"diff --git a/.rhiza/rhiza.mk b/.rhiza/rhiza.mk\n--- a/.rhiza/rhiza.mk\n+++ b/.rhiza/rhiza.mk\n@@ -169,7 +169,8 @@\n \n _apply-sync-schedule: ## (internal) apply RHIZA_SYNC_SCHEDULE override to GitHub Actions sync workflow\n \t@if [ \"$(RHIZA_SYNC_SCHEDULE)\" != \"0 0 * * 1\" ] && [ -f .github/workflows/rhiza_sync.yml ]; then \\\n-\t\tsed -i.bak \"s|cron: '[^']*'|cron: '$(RHIZA_SYNC_SCHEDULE)'|\" .github/workflows/rhiza_sync.yml && rm -f .github/workflows/rhiza_sync.yml.bak; \\\n+\t\tsed \"s|cron: '[^']*'|cron: '$(RHIZA_SYNC_SCHEDULE)'|\" .github/workflows/rhiza_sync.yml > .github/workflows/rhiza_sync.yml.tmp && \\\n+\t\tmv .github/workflows/rhiza_sync.yml.tmp .github/workflows/rhiza_sync.yml; \\\n \t\tprintf \"${BLUE}[INFO] Applied custom sync schedule: $(RHIZA_SYNC_SCHEDULE)${RESET}\\n\"; \\\n \tfi\n \n"}]},{"message":"The `summarise-sync` target already lists `install-uv` as a prerequisite on line 181, so the explicit `$(MAKE) install-uv` call inside the `else` branch is redundant. More importantly, the semicolon-separated recipe means a failure of `install-uv` here won't abort the recipe (unlike `&&`). If the intent is to ensure `install-uv` succeeds before continuing, use `&&` instead of `;`, consistent with how the `sync` target chains commands on line 164.","fixFiles":[{"filePath":".rhiza/rhiza.mk","diff":"diff --git a/.rhiza/rhiza.mk b/.rhiza/rhiza.mk\n--- a/.rhiza/rhiza.mk\n+++ b/.rhiza/rhiza.mk\n@@ -182,7 +182,7 @@\n \t@if [ -n \"$(IS_MOTHER_REPO)\" ]; then \\\n \t\tprintf \"${BLUE}[INFO] Skipping summarise-sync in rhiza repository (no template.yml by design)${RESET}\\n\"; \\\n \telse \\\n-\t\t$(MAKE) install-uv; \\\n+\t\t$(MAKE) install-uv && \\\n \t\t${UVX_BIN} \"rhiza==$(RHIZA_VERSION)\" summarise .; \\\n \tfi\n \n"}]},{"message":"The `available` parameter accepts `list[str] | None` but internally converts it to a list or None in line 55. Consider accepting `list[str]` with an empty list as default to simplify the API and remove the need for None-checking in client code. Alternatively, document clearly when `available` should be None versus an empty list.","fixFiles":[{"filePath":"src/jquantstats/exceptions.py","diff":"diff --git a/src/jquantstats/exceptions.py b/src/jquantstats/exceptions.py\n--- a/src/jquantstats/exceptions.py\n+++ b/src/jquantstats/exceptions.py\n@@ -38,8 +38,9 @@\n         jquantstats.exceptions.MissingDateColumnError: ...\n     \"\"\"\n \n-    def __init__(self, frame_name: str, column: str | None = None, available: list[str] | None = None) -> None:\n+    def __init__(self, frame_name: str, column: str | None = None, available: list[str] = []) -> None:\n         \"\"\"Initialize with the frame name and, optionally, the missing column and available columns.\"\"\"\n+        available = [] if available is None else list(available)\n         if column is None:\n             msg = f\"DataFrame '{frame_name}' is missing the required 'date' column.\"\n         else:\n@@ -52,7 +52,7 @@\n         super().__init__(msg)\n         self.frame_name = frame_name\n         self.column = column\n-        self.available = list(available) if available is not None else None\n+        self.available = available\n \n \n class InvalidCashPositionTypeError(JQuantStatsError, TypeError):\n"}]},{"message":"The parameter type `object` is too broad for `max_bps`. Consider using a more specific type hint like `int | float | object` or `typing.Any` to better communicate the expected input types while still allowing arbitrary values for the error message formatting.","fixFiles":[{"filePath":"src/jquantstats/exceptions.py","diff":"diff --git a/src/jquantstats/exceptions.py b/src/jquantstats/exceptions.py\n--- a/src/jquantstats/exceptions.py\n+++ b/src/jquantstats/exceptions.py\n@@ -15,7 +15,9 @@\n \n from __future__ import annotations\n \n+from typing import Any\n \n+\n class JQuantStatsError(Exception):\n     \"\"\"Base class for all JQuantStats domain errors.\"\"\"\n \n@@ -243,7 +244,7 @@\n         jquantstats.exceptions.InvalidMaxBpsError: max_bps must be a positive integer, got 0.\n     \"\"\"\n \n-    def __init__(self, max_bps: object) -> None:\n+    def __init__(self, max_bps: Any) -> None:\n         \"\"\"Initialize with the offending value.\"\"\"\n         super().__init__(f\"max_bps must be a positive integer, go...