feder-cr/Jobs_Applier_AI_Agent_AIHawk · error · ValueError

Could not extract section name from the response.

Error message

Could not extract section name from the response.

What it means

answer_question_textual_wide_range asks the LLM which resume section a question belongs to, then regex-matches the reply against a fixed list (Personal Info, Skills, Education, Experience Details, Projects, Availability, Salary Expectations, Certifications, Languages, Interests, Cover letter). If the model's free-text answer contains none of these keywords, the section cannot be determined and a ValueError is raised.

Source

Thrown at src/libs/llm_manager.py:573

            LANGUAGES: self._create_chain(prompts.languages_template),
            INTERESTS: self._create_chain(prompts.interests_template),
            COVER_LETTER: self._create_chain(prompts.coverletter_template),
        }

        prompt = ChatPromptTemplate.from_template(prompts.determine_section_template)
        chain = prompt | self.llm_cheap | StrOutputParser()
        raw_output = chain.invoke({QUESTION: question})
        output = self._clean_llm_output(raw_output)

        match = re.search(
            r"(Personal information|Self Identification|Legal Authorization|Work Preferences|Education "
            r"Details|Experience Details|Projects|Availability|Salary "
            r"Expectations|Certifications|Languages|Interests|Cover letter)",
            output,
            re.IGNORECASE,
        )
        if not match:
            raise ValueError("Could not extract section name from the response.")

        section_name = match.group(1).lower().replace(" ", "_")

        if section_name == "cover_letter":
            chain = chains.get(section_name)
            raw_output = chain.invoke(
                {
                    RESUME: self.resume,
                    JOB_DESCRIPTION: self.job_description,
                    COMPANY: self.job.company,
                }
            )
            output = self._clean_llm_output(raw_output)
            logger.debug(f"Cover letter generated: {output}")
            return output
        resume_section = getattr(self.resume, section_name, None) or getattr(
            self.job_application_profile, section_name, None
        )

View on GitHub (pinned to 79155b52fa)

Solutions

1. Retry the call: non-deterministic LLM output often matches on a second attempt.
2. Use a more instruction-following model (or lower temperature) for this chain so the reply names a valid section.
3. Extend the regex alternation in the prompt/validation to cover additional section names your resumes use.

Example fix

// before
answer = llm_manager.answer_question_textual_wide_range(q)
// after
for _ in range(3):
    try:
        answer = llm_manager.answer_question_textual_wide_range(q)
        break
    except ValueError:
        continue  # LLM reply didn't name a recognizable section

for attempt in range(3):
    try:
        return llm_manager.answer_question_textual_wide_range(question)
    except ValueError as e:
        if 'Could not extract section' not in str(e):
            raise
logger.warning('Section classification failed for: %s', question)

When it happens

Trigger: The LLM answers a screening question with wording that does not include any of the recognized section names (hallucinated or off-script reply), or the question is about something outside the enumerated sections.

Common situations: Flaky/creative LLM outputs, changing to a cheaper or different model that ignores instructions, prompt template changes, or non-English model responses.

Related errors

• No numbers found in the string
• Unsupported model type: {llm_model_type}
• Section '{section_name}' not found in either resume or job_a
• Chain not defined for section '{section_name}'
• Failed to get a response from the model after multiple attem

AI-assisted analysis of feder-cr/Jobs_Applier_AI_Agent_AIHawk@79155b52fa (2026-08-28). Data as JSON: /api/errors/65746c2ae15927d8. Report an issue: GitHub.