icl

[huangnie2020]

No description provided.

[gemini-code-assist]

Code Review

This pull request refactors the LongContext-ICL-Annotation pipeline to support Ascend NPU environments, batch processing, and multi-service deployment. It introduces robust API client error handling, optimized prompt construction, and retry mechanisms for failed requests. The review feedback highlights several critical and medium-severity improvements, including PEP 8 style guide compliance (such as using is not None and implicit boolean checks), fixing incorrect type hints, and refactoring fragile string parsing logic to be more robust.

[gemini-code-assist]

Any transient network error, rate limit, or temporary API issue will raise an exception and crash the entire batch run, losing all progress. It is highly recommended to handle these exceptions gracefully by logging the error and returning None, False, allowing the retry logic in main_batch.py to handle it later.

        except APIConnectionError as e:
            print(f"连接失败: {e}")
            print("请检查 base_url 是否配置正确，或者网络是否正常。")
            return None, False
        except AuthenticationError as e:
            print(f"鉴权失败: {e}")
            print("请检查 API Key 是否填写正确。")
            return None, False
        except APIError as e:
            error = str(e)
            if "inappropriate" in error:
                return '<output_answer>Output data may contain inappropriate content</output_answer>', False
            else:
                print(f"API错误: {e}")
                return None, False
        except BadRequestError as e:
            error = str(e)
            if "inappropriate" in error:
                return '<output_answer>Input data may contain inappropriate content</output_answer>', False
            else:
                print(f"请求错误: {e}")
                return None, False
        except Exception as e:
            error = str(e)
            if "inappropriate" in error:
                return '<output_answer>Output data may contain inappropriate content</output_answer>', False
            else:
                print(f"未知错误: {e}")
                return None, False

[gemini-code-assist]

According to PEP 8, comparisons to singletons like None should always be done with is or is not, never the equality operators.

Suggested change

	if content != None:
	if content is not None:

References

1. Comparisons to singletons like None should always be done with is or is not, never the equality operators. ^(link)

[gemini-code-assist]

According to PEP 8, comparisons to boolean values should be written implicitly (e.g., if not state: instead of if state == False:).

Suggested change

	if state == False:
	if not state:

References

1. Comparisons to True or False should be written implicitly, never using equality operators. ^(link)

[gemini-code-assist]

According to PEP 8, comparisons to boolean values should be written implicitly (e.g., if not state: instead of if state == False:).

Suggested change

	if state == False:
	if not state:

References

1. Comparisons to True or False should be written implicitly, never using equality operators. ^(link)

[gemini-code-assist]

According to PEP 8, comparisons to singletons like None should always be done with is or is not, never the equality operators.

Suggested change

	if answer == None:
	return None, False
	if answer is None:
	return None, False

References

1. Comparisons to singletons like None should always be done with is or is not, never the equality operators. ^(link)

[gemini-code-assist]

According to PEP 8, comparisons to boolean values should be written implicitly (e.g., not data['state'] instead of data['state'] == False).

Suggested change

	if data['state'] == False or data_len < min_len or data_len > max_len:
	if not data['state'] or data_len < min_len or data_len > max_len:

References

1. Comparisons to True or False should be written implicitly, never using equality operators. ^(link)

[gemini-code-assist]

Checking if a dictionary is empty can be done more pythonically by evaluating the dictionary directly (e.g., if not checks: instead of if len(checks.values()) == 0:).

Suggested change

	if len(checks.values()) == 0:
	if not checks:

[gemini-code-assist]

The variable name indexs is misspelled (should be indices or indexes). Additionally, checking membership in checks directly is more idiomatic and direct than calling checks.keys().

Suggested change

	indexs = [ i for i, item in enumerate(test_samples) if item['id'] in check_ids ]
	for k in range(0, len(indexs), req_concurrency_size):
	sample_idxs = indexs[k : k + req_concurrency_size]
	indices = [ i for i, item in enumerate(test_samples) if item['id'] in checks ]
	for k in range(0, len(indices), req_concurrency_size):
	sample_idxs = indices[k : k + req_concurrency_size]