From 5219fc65754e28d2f9860eaaa8459395a6fb8d61 Mon Sep 17 00:00:00 2001
From: Alfonso Lozana <57018030+AlfonsoLozana@users.noreply.github.com>
Date: Wed, 29 Jul 2026 15:07:34 +0200
Subject: [PATCH] feat: add reader overrides (#2317)

* feat: add reader overrides

* fix: fix disable readers

* fix: change override to register

* fix: control UnidentifiedImageError

* Revert "fix: control UnidentifiedImageError"

This reverts commit df2cc8cceec81eeb61afba9aad4679eb5ebdd968.

* fix: fallback resolved reader

* chore: remove not necessary var
---
 private_gpt/components/ingest/parse_component.py | 2 ++
 private_gpt/components/readers/registry.py       | 6 ++++++
 2 files changed, 8 insertions(+)

diff --git a/private_gpt/components/ingest/parse_component.py b/private_gpt/components/ingest/parse_component.py
index 89579705..791e5783 100644
--- a/private_gpt/components/ingest/parse_component.py
+++ b/private_gpt/components/ingest/parse_component.py
@@ -134,8 +134,10 @@ class ParseComponent:
             raise InvalidFileError(errors=[IngestionParseErrors.PARSING_FAILURE]) from e
         except Exception as e:
             logger.error("Error loading file: %s", e, exc_info=True)
+
             converted_fallback = convert_unsupported_file_as_fallback(file_info)
             if converted_fallback:
+                resolved_reader = self._resolve_reader(converted_fallback.extension)
                 if notification:
                     notification(
                         percentage=0,
diff --git a/private_gpt/components/readers/registry.py b/private_gpt/components/readers/registry.py
index 89a180b4..efa51ec2 100644
--- a/private_gpt/components/readers/registry.py
+++ b/private_gpt/components/readers/registry.py
@@ -25,6 +25,12 @@ _DEFAULT_EXTENSION_READERS: dict[str, list[str]] = {
 }
 
 
+def register_extension_readers(extension: str, reader_names: list[str]) -> None:
+    normalized_extension = _normalize_extension(extension)
+    normalized_reader_names = [_normalize_reader_name(n) for n in reader_names]
+    _DEFAULT_EXTENSION_READERS[normalized_extension] = normalized_reader_names
+
+
 def _normalize_reader_name(name: str) -> str:
     normalized = name.strip().lower()
     if not normalized: