transcript-viewer 0.8.0__tar.gz → 0.8.1__tar.gz

{transcript_viewer-0.8.0 → transcript_viewer-0.8.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: transcript-viewer
-Version: 0.8.0
+Version: 0.8.1
 Summary: Browse agent transcripts in a local, dependency-free web viewer.
 License: MIT
 Requires-Python: >=3.12

{transcript_viewer-0.8.0 → transcript_viewer-0.8.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "transcript-viewer"
-version = "0.8.0"
+version = "0.8.1"
 description = "Browse agent transcripts in a local, dependency-free web viewer."
 readme = "README.md"
 requires-python = ">=3.12"

{transcript_viewer-0.8.0 → transcript_viewer-0.8.1}/src/transcript_viewer/cli.py

@@ -49,7 +49,12 @@ def cmd_view(args: argparse.Namespace) -> int:
         # An explicit path that holds nothing is a mistake worth reporting; an
         # empty library is not, so this only guards the argument.
         if not entries:
-            print(f"transcript-viewer: nothing convertible in {path}", file=sys.stderr)
+            print(
+                f"transcript-viewer: no transcript in {path}. "
+                f"{corpus.describe_contents(path)}, and none of it is a log "
+                f"this can read.",
+                file=sys.stderr,
+            )
             return 1
     else:
         # Only what has been added deliberately. Scanning a machine because the

{transcript_viewer-0.8.0 → transcript_viewer-0.8.1}/src/transcript_viewer/corpus.py

@@ -9,6 +9,7 @@ something actually asks for that trajectory.
 from __future__ import annotations
 
 import hashlib
+from collections import Counter
 import re
 import json
 from dataclasses import asdict, dataclass
@@ -367,3 +368,32 @@ def load(path: Path | None = None) -> list[Entry]:
         except TypeError:
             continue
     return entries
+
+
+def describe_contents(root: Path, limit: int = 4000) -> str:
+    """What is inside a path, said in a sentence.
+
+    Reached only when nothing in it converted. "Nothing convertible" on its own
+    invites the reader to doubt the tool rather than the file — one real archive
+    held 1,952 files and an audit that referenced a transcript which had not
+    been shipped with it, and the answer gave no way to tell that from a bug.
+    """
+    if root.is_file():
+        return f"{root.name} ({root.stat().st_size:,} bytes)"
+
+    kinds: Counter[str] = Counter()
+    total = 0
+    for item in root.rglob("*"):
+        if not item.is_file():
+            continue
+        total += 1
+        if total > limit:
+            break
+        kinds[item.suffix.lower() or "no extension"] += 1
+
+    if not total:
+        return "it is empty"
+
+    common = ", ".join(f"{n:,} {ext}" for ext, n in kinds.most_common(4))
+    more = "+" if total > limit else ""
+    return f"{total:,}{more} files inside — {common}"

{transcript_viewer-0.8.0 → transcript_viewer-0.8.1}/src/transcript_viewer/viewer.py

@@ -27,6 +27,7 @@ from urllib.parse import parse_qs, unquote, urlparse
 from atif_make.atif import ContentPart, Trajectory
 from . import corpus
 from atif_make.container import is_container, open_container
+from atif_make.detect import FORMATS
 from atif_make.convert import convert
 from transcript_viewer.corpus import Entry, scan
 
@@ -615,9 +616,19 @@ class _Handler(BaseHTTPRequestHandler):
             self._json({"error": str(exc)}, 400)
             return
         if not found:
-            # Nothing usable in it — do not keep the copy around.
+            # Nothing usable in it — do not keep the copy around. Say what was
+            # in it on the way out: an answer of "nothing convertible" with
+            # nothing behind it reads as the tool failing rather than the file
+            # not holding what was hoped.
+            inside = corpus.describe_contents(target)
             shutil.rmtree(home, ignore_errors=True)
-            self._json({"error": f"nothing convertible in {name}"}, 415)
+            self._json(
+                {
+                    "error": f"No transcript in {name}. {inside}, and none of it "
+                    f"is a log this can read ({', '.join(FORMATS)})."
+                },
+                415,
+            )
             return
 
         with self.lock:

{transcript_viewer-0.8.0 → transcript_viewer-0.8.1}/tests/test_cli.py

@@ -65,7 +65,11 @@ def test_a_path_with_nothing_convertible_reports_rather_than_serving(
 ):
     (tmp_path / "notes.txt").write_text("nothing here")
     assert cli.main([str(tmp_path)]) == 1
-    assert "nothing convertible" in capsys.readouterr().err
+    said = capsys.readouterr().err
+    assert "no transcript" in said
+    # It must also say what it looked at: a bare refusal reads as the tool
+    # failing rather than the path not holding what was hoped for.
+    assert "1 file" in said and ".txt" in said
     assert not served
 
 

{transcript_viewer-0.8.0 → transcript_viewer-0.8.1}/tests/test_corpus.py

@@ -496,3 +496,36 @@ def test_pieces_unpacked_beside_a_download_are_used_as_they_are(tmp_path):
     entries = corpus.scan([tmp_path])
     assert len(entries) == 3
     assert all(str(beside) in e.path for e in entries)
+
+
+def test_a_path_that_converts_to_nothing_still_says_what_was_in_it(tmp_path):
+    """"Nothing convertible" alone reads as the tool failing, not the file."""
+    (tmp_path / "a").mkdir()
+    (tmp_path / "a" / "notes.md").write_text("# hello")
+    (tmp_path / "a" / "data.json").write_text('{"not": "a transcript"}')
+    (tmp_path / "a" / "run.py").write_text("print(1)")
+
+    said = corpus.describe_contents(tmp_path)
+
+    assert "3 files" in said
+    for ext in (".md", ".json", ".py"):
+        assert ext in said, f"{ext} should be named"
+
+
+def test_describing_a_single_file_names_it(tmp_path):
+    path = tmp_path / "one.jsonl"
+    path.write_text("{}\n")
+    said = corpus.describe_contents(path)
+    assert "one.jsonl" in said and "bytes" in said
+
+
+def test_describing_an_empty_directory_says_so(tmp_path):
+    assert corpus.describe_contents(tmp_path) == "it is empty"
+
+
+def test_describing_a_huge_tree_does_not_read_all_of_it(tmp_path):
+    """A 600 MB archive should not be walked twice to explain itself."""
+    for i in range(60):
+        (tmp_path / f"f{i}.bin").write_text("x")
+    said = corpus.describe_contents(tmp_path, limit=10)
+    assert "+" in said, "it should say the count is a floor, not a total"

{transcript_viewer-0.8.0 → transcript_viewer-0.8.1}/uv.lock

@@ -388,7 +388,7 @@ wheels = [
 
 [[package]]
 name = "transcript-viewer"
-version = "0.8.0"
+version = "0.8.1"
 source = { editable = "." }
 dependencies = [
     { name = "atif-make" },