Canonical prefix-code assignment
Codelith is a Windows AI coding client for examining how code lengths become concrete bit strings. Discuss symbol ordering with a compatible model, review its proposed edits and test the table for prefix conflicts before integration.
def canonical_codes(lengths):
if type(lengths) is not dict or len(lengths) > 256:
raise ValueError("codes.input")
if any(type(s) is not str or not s or type(n) is not int or
not 1 <= n <= 16 for s,n in lengths.items()):
raise ValueError("codes.length")
result, code, previous = {}, 0, 0
for symbol, length in sorted(lengths.items(), key=lambda p:(p[1],p[0])):
code <<= length-previous
if code >= 1 << length:
raise ValueError("codes.oversubscribed")
result[symbol] = format(code, f"0{length}b")
code += 1
previous = length
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 empty and incomplete tables, sort equal lengths by the original Unicode strings and reject oversubscribed or invalid lengths without normalizing symbols. |
def canonical_codes(lengths):
if type(lengths) is not dict or len(lengths) > 256:
raise ValueError("codes.input")
if any(type(s) is not str or not s or type(n) is not int or
not 1 <= n <= 16 for s,n in lengths.items()):
raise ValueError("codes.length")
result, code, previous = {}, 0, 0
for symbol, length in sorted(lengths.items(), key=lambda p:(p[1],p[0])):
code <<= length-previous
if code >= 1 << length:
raise ValueError("codes.oversubscribed")
result[symbol] = format(code, f"0{length}b")
code += 1
previous = length
return resultMake an encoding-table change reproducible
When two implementations build different encoding tables from the same description, debugging needs more than another sample output. Codelith gives that investigation a Windows AI coding workspace: open the project, put the competing results in the conversation and ask your connected model to inspect the source of the difference before suggesting changes.
One suitable task is canonical prefix-code assignment from symbol lengths. Describe the ordering rule for equal lengths and whether the original Unicode strings must remain unchanged. Include the consumer of the table so the proposal accounts for the format it must read. The assignment example is a way to frame your code review, rather than an encoding tool bundled with the workspace.
Ask for a focused proposal that makes the chosen ordering explicit. Review empty and incomplete tables separately from oversubscribed lengths, then run your format fixtures in the project's own environment. Compare exact bit strings and symbol associations. An orderly table can still disagree with its consumer. Keep expected serialization visible during review.
Codelith supports model connections through OpenAI-compatible or Anthropic endpoints. An unrestricted AI model is a selectable possibility when your provider offers it, with credentials as needed. The selected server and model govern context, tools and restrictions. Your fixtures and review establish whether the proposed encoding change meets the application contract.
Questions for an encoding-table change
- What evidence helps explain two different tables?
Provide the same symbol-length input, each resulting table and the documented ordering convention. Identify where the application writes or reads the table. The model can then discuss a concrete discrepancy instead of guessing which output the format expects.
- Should Unicode symbols be normalized before sorting?
Only if the format explicitly requires that transformation. The illustrative task preserves original strings and orders them accordingly. State the convention in the request, then verify that the proposed code does not silently merge or alter distinct symbols.
- Are incomplete and oversubscribed tables equivalent errors?
They describe different conditions, and your format determines which it accepts. Include separate fixtures and expected outcomes. Ask the proposal to preserve that distinction rather than treating any unused code space as evidence that a table is invalid.
- How do I check compatibility with an existing reader?
Run representative table fixtures through your project's writer and reader, comparing the exact associations required by the format. Review both paths if the proposal changes ordering. A successful source edit alone does not establish compatibility with previously stored data.
Codelith · Updated 2026-10-25