Denying empty text going to citation_prompt (#998)

jamesbraza · Copilot · web-flow · commit 2bea09183caf · 2025-07-08T17:47:40.000-07:00
Co-authored-by: Copilot &lt;175728472+Copilot@users.noreply.github.com&gt;
diff --git a/paperqa/docs.py b/paperqa/docs.py
@@ -283,7 +283,7 @@ async def aadd(  # noqa: PLR0912
                 use_block_parsing=parse_config.pdfs_use_block_parsing,
                 parse_pdf=parse_config.parse_pdf,
             )
-            if not texts:
+            if not texts or not texts[0].text.strip():
                 raise ValueError(f"Could not read document {path}. Is it empty?")
             result = await llm_model.call_single(
                 messages=[

Original file line number	Diff line number	Diff line change
`@@ -283,7 +283,7 @@ async def aadd( # noqa: PLR0912`
`283`	`283`	`use_block_parsing=parse_config.pdfs_use_block_parsing,`
`284`	`284`	`parse_pdf=parse_config.parse_pdf,`
`285`	`285`	`)`
`286`		`- if not texts:`
	`286`	`+ if not texts or not texts[0].text.strip():`
`287`	`287`	`raise ValueError(f"Could not read document {path}. Is it empty?")`
`288`	`288`	`result = await llm_model.call_single(`
`289`	`289`	`messages=[`