From 31a85668d16462e9fde6ccb920b928f86605e6c1 Mon Sep 17 00:00:00 2001
From: xumian520 <126989134+xumian520@users.noreply.github.com>
Date: Sun, 26 Jul 2026 14:59:35 +0800
Subject: [PATCH 1/2] test: read generated outputs as UTF-8 on Windows

---
 tests/test_merge_chunks_validation.py | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/tests/test_merge_chunks_validation.py b/tests/test_merge_chunks_validation.py
index b2d532591..6530dbab2 100644
--- a/tests/test_merge_chunks_validation.py
+++ b/tests/test_merge_chunks_validation.py
@@ -31,7 +31,7 @@ def test_merge_chunks_skips_chunk_with_path_escape_id(tmp_path, monkeypatch, cap
 
     _run_merge(monkeypatch, ["graphify", "merge-chunks", str(good), str(bad), "--out", str(out)])
 
-    merged = json.loads(out.read_text())
+    merged = json.loads(out.read_text(encoding="utf-8"))
     assert {n["id"] for n in merged["nodes"]} == {"pkg.mod.good"}
     captured = capsys.readouterr()
     assert "skipping invalid chunk" in captured.err
@@ -48,7 +48,7 @@ def test_merge_chunks_fails_closed_when_every_chunk_is_invalid(tmp_path, monkeyp
         _run_merge(monkeypatch, ["graphify", "merge-chunks", str(bad), "--out", str(out)])
 
     assert exc.value.code == 1
-    assert json.loads(out.read_text()) == {"previous": "semantic result"}
+    assert json.loads(out.read_text(encoding="utf-8")) == {"previous": "semantic result"}
     err = capsys.readouterr().err
     assert "skipping invalid chunk" in err
     assert "no valid chunks to merge" in err
@@ -62,7 +62,7 @@ def test_merge_chunks_accepts_valid_empty_chunk(tmp_path, monkeypatch):
 
     _run_merge(monkeypatch, ["graphify", "merge-chunks", str(empty), "--out", str(out)])
 
-    merged = json.loads(out.read_text())
+    merged = json.loads(out.read_text(encoding="utf-8"))
     assert merged["nodes"] == []
     assert merged["edges"] == []
 
@@ -87,7 +87,7 @@ def test_merge_chunks_fails_closed_on_unmatched_glob(tmp_path, monkeypatch, caps
         _run_merge(monkeypatch, ["graphify", "merge-chunks", unmatched, "--out", str(out)])
 
     assert exc.value.code == 1
-    assert json.loads(out.read_text()) == {"previous": True}
+    assert json.loads(out.read_text(encoding="utf-8")) == {"previous": True}
     err = capsys.readouterr().err
     assert "skipping invalid chunk" in err
     assert "no valid chunks to merge" in err
@@ -102,7 +102,7 @@ def test_merge_chunks_accepts_synonym_file_type(tmp_path, monkeypatch):
                "edges": [], "hyperedges": []})
     out = tmp_path / "merged.json"
     _run_merge(monkeypatch, ["graphify", "merge-chunks", str(c), "--out", str(out)])
-    merged = json.loads(out.read_text())
+    merged = json.loads(out.read_text(encoding="utf-8"))
     assert {n["id"] for n in merged["nodes"]} == {"pkg.readme", "pkg.tool"}
 
 
@@ -114,7 +114,7 @@ def test_merge_chunks_accepts_unicode_id(tmp_path, monkeypatch):
                "edges": [], "hyperedges": []})
     out = tmp_path / "merged.json"
     _run_merge(monkeypatch, ["graphify", "merge-chunks", str(c), "--out", str(out)])
-    merged = json.loads(out.read_text())
+    merged = json.loads(out.read_text(encoding="utf-8"))
     assert {n["id"] for n in merged["nodes"]} == {"mod_处理数据"}
 
 
@@ -144,7 +144,7 @@ def test_merge_chunks_merges_valid_chunks(tmp_path, monkeypatch):
 
     _run_merge(monkeypatch, ["graphify", "merge-chunks", str(c0), str(c1), "--out", str(out)])
 
-    merged = json.loads(out.read_text())
+    merged = json.loads(out.read_text(encoding="utf-8"))
     assert {n["id"] for n in merged["nodes"]} == {"a", "b"}
     assert merged["input_tokens"] == 17
     assert merged["output_tokens"] == 8

From f26d1a9ab3cb02786188c681ca0fb9ebab5653c6 Mon Sep 17 00:00:00 2001
From: xumian520 <126989134+xumian520@users.noreply.github.com>
Date: Sun, 26 Jul 2026 15:01:32 +0800
Subject: [PATCH 2/2] test: read pipeline outputs as UTF-8 on Windows

Update file reading to specify UTF-8 encoding for JSON and HTML.
---
 tests/test_pipeline.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tests/test_pipeline.py b/tests/test_pipeline.py
index ce6055d8b..2cf7ae91d 100644
--- a/tests/test_pipeline.py
+++ b/tests/test_pipeline.py
@@ -71,7 +71,7 @@ def run_pipeline(tmp_path: Path) -> dict:
     json_path = tmp_path / "graph.json"
     to_json(G, communities, str(json_path))
     assert json_path.exists()
-    data = json.loads(json_path.read_text())
+    data = json.loads(json_path.read_text(encoding="utf-8"))
     assert "nodes" in data and "links" in data
     assert all("community" in n for n in data["nodes"])
 
@@ -79,7 +79,7 @@ def run_pipeline(tmp_path: Path) -> dict:
     html_path = tmp_path / "graph.html"
     to_html(G, communities, str(html_path), community_labels=labels)
     assert html_path.exists()
-    html = html_path.read_text()
+    html = html_path.read_text(encoding="utf-8")
     assert "vis-network" in html
     assert "RAW_NODES" in html