infiniflow/ragflow · error · Exception

PDF generation failed: {str(e)}

Error message

PDF generation failed: {str(e)}

What it means

Catch-all wrapper thrown by DocGenerator._generate_pdf (agent/component/docs_generator.py): any exception from the pandoc PDF conversion (engine invocation, header/font args, overlay step) is re-raised as Exception('PDF generation failed: <original message>'). The root cause is in the appended inner text — always read the suffix, not this headline.

Source

Thrown at agent/component/docs_generator.py:614

            try:
                file_path, _ = self._generate_pandoc_binary_output(
                    content,
                    "pdf",
                    "pdf",
                    include_timestamp_in_body=False,
                    extra_args=[
                        "--standalone",
                        f"--pdf-engine={engine}",
                        f"--include-in-header={header_path}",
                        *self._get_pdf_font_args(),
                    ],
                )
            finally:
                if os.path.exists(header_path):
                    os.remove(header_path)
            return self._apply_pdf_overlay(file_path)
        except Exception as e:
            raise Exception(f"PDF generation failed: {str(e)}")

    def _generate_docx(self, content: str) -> tuple[str, bytes]:
        try:
            file_path, _ = self._generate_pandoc_binary_output(
                content,
                "docx",
                "docx",
                include_timestamp_in_body=False,
                extra_args=["--standalone"],
            )
            return self._decorate_docx(file_path)
        except Exception as e:
            raise Exception(f"DOCX generation failed: {str(e)}")

    def _generate_txt(self, content: str) -> tuple[str, bytes]:
        try:
            return self._generate_pandoc_text_output(content, "plain", "txt")
        except Exception as e:

View on GitHub (pinned to 554fb1133a)

Solutions

  1. Read the text after 'PDF generation failed:' — it names the real fault (e.g. 'Font Noto Sans CJK SC not found')
  2. Install the font named in the inner error (fonts-noto-cjk) or the missing engine per error 154 guidance
  3. Reproduce outside the agent: echo content | pandoc -o out.pdf --pdf-engine=xelatex -V mainfont='Noto Sans CJK SC'
  4. If content is the problem, sanitize markdown (escape LaTeX-special chars like $, %, &, {}) or switch to docx/html output

Example fix

# before: content with raw LaTeX specials
content = "Cost is 100% of $50 & rising"

# after: escape or wrap
content = "Cost is 100\\% of \$50 \& rising"
Defensive patterns

Strategy: try-catch

Validate before calling

import shutil

def pdf_prerequisites_ok():
    return shutil.which('xelatex') and shutil.which('pandoc')

# plus: verify fonts exist, e.g. fc-list | grep -i 'Noto Sans CJK'

Try / catch

try:
    file_path, file_bytes = generator._generate_pdf(content)
except Exception as e:
    inner = str(e).removeprefix('PDF generation failed: ')
    logger.error('PDF generation root cause: %s', inner)
    if 'not found' in inner and 'font' in inner.lower():
        hint = 'Install fonts-noto-cjk'
    elif 'pandoc' in inner.lower():
        hint = 'Install pandoc/pypandoc-binary'
    else:
        hint = 'Reproduce with pandoc --pdf-engine=xelatex'
    raise RuntimeError(f'{inner} ({hint})') from e

Prevention

When it happens

Trigger: Running DocGenerator with output_format 'pdf' where pandoc/xelatex fails: missing fonts specified by _get_pdf_font_args (Noto Sans CJK SC), malformed markdown that xelatex rejects, pandoc not installed (pypandoc raises OSError), or the _apply_pdf_overlay step failing. The try wraps the whole pandoc call plus overlay.

Common situations: Containers with xelatex but without the CJK fonts; pandoc version mismatches; special characters/unbalanced braces in content breaking LaTeX; reportlab overlay errors (missing STSong-Light CID font).

Related errors

AI-assisted analysis of infiniflow/ragflow@554fb1133a (2026-08-15). Data as JSON: /api/errors/328048f5f4b6244e. Report an issue: GitHub.