TickFable supports AI-assisted review of duration parsing in a project folder. Connect a compatible model, discuss the exact token grammar and inspect rejection cases before adopting its suggested patch.
import re
def duration_ms(text):
if not isinstance(text, str):
raise ValueError("duration.text")
match = re.fullmatch(r"(0|[1-9][0-9]{0,8})(ms|s|m|h)", text)
if not match:
raise ValueError("duration.syntax")
result = int(match[1]) * {"ms": 1, "s": 1000, "m": 60000, "h": 3600000}[match[2]]
if result > 86400000:
raise ValueError("duration.range")
return result| Review before applying | Compare actual behavior with the expected result. If a check fails, investigate it before applying more changes. |
|---|---|
| Connection | Base URL, credentials and a reachable server. |
| Project | Check zero and the day boundary, rejecting leading zeros, fractions, spaces and unknown units. |
import re
def duration_ms(text):
if not isinstance(text, str):
raise ValueError("duration.text")
match = re.fullmatch(r"(0|[1-9][0-9]{0,8})(ms|s|m|h)", text)
if not match:
raise ValueError("duration.syntax")
result = int(match[1]) * {"ms": 1, "s": 1000, "m": 60000, "h": 3600000}[match[2]]
if result > 86400000:
raise ValueError("duration.range")
return resultA setting that accepts a duration needs a policy users and callers can rely on. TickFable provides a Windows AI coding workspace for working through that policy in source. Open the project, identify the setting's parser and consumer, and explain which inputs should succeed before asking the connected model for a proposed change.
One example task accepts an ASCII integer followed by ms, s, m or h, converts it to milliseconds and caps the result at one day. Discuss whether spaces, fractions and leading zeros belong in the accepted grammar. Those decisions connect the parser to product behavior; they should not emerge accidentally from whichever expression the model happens to propose.
Review the suggested edit against a short acceptance table from your own requirements. Check zero, the day boundary and one value beyond it, then exercise the application's message for rejected input. Run the checks in your existing environment. The useful outcome is a consistent setting contract from text entry through stored value, not merely a parser that recognizes the happy path.
TickFable connects to supported OpenAI-compatible or Anthropic endpoints, with credentials where required. An unrestricted AI model may be one available selection. Its modification does not define your grammar or establish validation accuracy; context capacity, tools and restrictions remain determined by the chosen model and server.
Provide a small table of exact strings and expected outcomes, including capitalization and whitespace. Include the units your setting supports. This makes the proposed edit easier to inspect and avoids relying on an informal description such as human-readable duration.
Ask the model to inspect how the parsed quantity becomes the unit consumed by the application. Your review should confirm that the limit applies to the converted value. Exercise equivalent boundary values in different supported units using your project tests.
Only if that behavior is part of your requirement. The illustrative strict grammar rejects spaces. Tell the conversation whether your interface intentionally normalizes input before parsing, then review that boundary so the same value is not treated inconsistently by different callers.
Inspect how the caller stores and uses the converted duration and how it reports rejection. A correct parser can still feed the wrong unit into a timer. Keep that consumer in the task and verify the observable behavior in your application.
TickFable · Updated 2026-10-25