From 0210208972bb81bcd21aa1bc7f3a04b7775d8e63 Mon Sep 17 00:00:00 2001
From: Keoni Gandall <koeng101@gmail.com>
Date: Sat, 14 Dec 2024 11:07:40 -0800
Subject: [PATCH] fix iso-8859-1

---
 lib/bio/uniprot/uniprot.go | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/lib/bio/uniprot/uniprot.go b/lib/bio/uniprot/uniprot.go
index 029a63f..63a367a 100644
--- a/lib/bio/uniprot/uniprot.go
+++ b/lib/bio/uniprot/uniprot.go
@@ -28,6 +28,7 @@ import (
 	"io"
 	"net/http"
 	"net/url"
+	"strings"
 )
 
 // Decoder decodes XML elements2
@@ -67,6 +68,14 @@ type Parser struct {
 
 func NewParser(r io.Reader) *Parser {
 	decoder := xml.NewDecoder(r)
+	// Oddly enough, the uniref datasets use iso-8859-1, not UTF-8. So we need
+	// to incorporate this decoder charset reader.
+	decoder.CharsetReader = func(charset string, input io.Reader) (io.Reader, error) {
+		if strings.ToLower(charset) == "iso-8859-1" {
+			return input, nil // ISO-8859-1 bytes can be read directly as UTF-8
+		}
+		return nil, fmt.Errorf("unsupported charset: %s", charset)
+	}
 	return &Parser{decoder: decoder}
 }