Replay text-encoded files in gr.load_chat history as text, not as images

[abidlabs]

Fixes #11331

Root cause

With the default gr.load_chat(..., file_types="text_encoded"), pasting a large prompt turns it into a text-file attachment. format_conversation() handles that correctly on the turn it is sent — it reads text-encoded files and appends their contents to the prompt as text.

But when it walks the history, it turned every file content part into an image regardless of type:

elif content["type"] == "file":
    new_content.append({"type": "image_url", "image_url": {"url": encode_url_or_file_to_base64(...)}})

So the pasted-text turn goes out fine, and then every message after it re-sends that .txt base64-encoded as an image — the payload literally contains "url": "data:text/plain;base64,...". That is the request vllm serve rejects with "model is not multimodal", which matches the reported sequence exactly: paste long text (works) → type something short (fails).

It also explains why file_types=[] made the error go away: no attachment, so nothing to replay.

Fix

Text-encoded files in the history are inlined as text, the same as when they were first sent. Images still go out as image_url. The extension check and the ## filename\ncontents formatting are now shared between the two paths so they cannot drift again, and the check also excludes http(s) URLs, which Path(...).read_text() could not have handled anyway.

Verification

$ python -m pytest test/test_external.py -k format_conversation -q
1 passed

test_format_conversation_replays_text_files_as_text is new and fails on main:

E  AssertionError: assert [{'type': 'im...Byb21wdA=='}}] == [{'type': 'te...sted prompt'}]

It also pins that a .png in the history is still sent as image_url. Full test/test_external.py: the 3 pre-existing failures are unchanged (test_load_chat_*, which need openai, not installed in my env). TestLoadInterface::test_multiple_spaces_one_private flaked in one batch run and passes on its own — it hits real Spaces over the network.

End-to-end, with the demo below (a local stand-in for vllm serve that rejects image content the way a non-multimodal model does):

before:  payload content types: ['image_url', 'text', 'text']
         BadRequestError: Error code: 400 - ... 'Model Qwen/Qwen3-8B is not multimodal, but multimodal input was provided.'
after:   payload content types: ['text', 'text', 'text']
         response: ok

Minimal demo (not committed)

from pathlib import Path
from gradio.external import format_conversation

pasted = Path("pasted_text.txt")
pasted.write_text("a very long pasted prompt\n" * 50)

# what the Chatbot holds after a turn where a long prompt was pasted
history = [
    {"role": "user", "content": [{"type": "file", "file": {"path": str(pasted)}}]},
    {"role": "assistant", "content": [{"type": "text", "text": "ok"}]},
]

conversation = format_conversation(history, "and now a short one")
print([part["type"] for m in conversation for part in m["content"]])

The full demo file (which also starts a fake OpenAI-compatible server and drives a real gr.load_chat) was kept untracked and is not part of this PR.

🤖 Generated with Claude Code

[gradio-pr-bot]

🪼 branch checks and previews

•	Name	Status	URL
	Spaces	ready!	Spaces preview
	Website	ready!	Website preview
🦄	Changes	detected!	Details

---

Install Gradio from this PR

pip install https://huggingface.co/buckets/gradio/pypi-previews/resolve/607eb88ee96c9fff301565e0a2b01688bac13aaf/gradio-6.23.1-py3-none-any.whl

Install Gradio Python Client from this PR

pip install "gradio-client @ git+https://github.com/gradio-app/gradio@607eb88ee96c9fff301565e0a2b01688bac13aaf#subdirectory=client/python"

Import Gradio JS Client from this PR via CDN

import { Client } from "https://huggingface.co/buckets/gradio/npm-previews/resolve/607eb88ee96c9fff301565e0a2b01688bac13aaf/browser.js";

[gradio-pr-bot]

🦄 change detected

This Pull Request includes changes to the following packages.

Package	Version
gradio	patch

---

• Keep text-encoded files as text when replaying gr.load_chat history

---

Something isn't right?

• Maintainers can change the version label to modify the version bump.
• If the bot has failed to detect any changes, or if this pull request needs to update multiple packages to different versions or requires a more comprehensive changelog entry, maintainers can update the changelog file directly.

[abidlabs]

Folded into #13742, which now carries both chat-history fixes (this one and #10823) — the two touch the same area, so they are easier to review together. No content lost: the commit here was cherry-picked onto that branch unchanged.

Closing in favour of #13742.