Update Eskin and Bill Black

.github/workflows/ci.yml

@@ -32,13 +32,23 @@ jobs:
           python -m pip install flit
           flit install -s
 
-      - name: Test with pytest
+      - name: Test with pytest (non-web)
+        if: ${{ matrix.python-version != '3.14' }}
         run: >
           pytest -v -ra
           --doctest-modules
           --cov=pyabc2 --cov-report xml --cov-report term-missing
+          -m "not web"
           -n auto
 
+      - name: Test with pytest (full, including web)
+        if: ${{ matrix.python-version == '3.14' }}
+        run: >
+          pytest -v -ra
+          --doctest-modules
+          --cov=pyabc2 --cov-report xml --cov-report term-missing
+          -n auto --dist loadgroup
+
       - name: Check type annotations
         run: |
           mypy --non-interactive .

docs/changes.md

@@ -1,5 +1,9 @@
 # Release notes
 
+## v0.1.4 (unreleased)
+
+* Updates for Eskin and Bill Black ({pull}`112`)
+
 ## v0.1.3 (2026-04-17)
 
 * Improve styling of abcjs containers, pandas dataframes, and ipywidgets

docs/examples/sources.ipynb

@@ -402,7 +402,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "Tune(df.query(\"group == 'jigs'\").iloc[0].abc)"
+    "Tune(df.query(\"group == 'Jig A-G'\").iloc[0].abc)"
    ]
   },
   {

pyabc2/sources/bill_black.py

@@ -8,60 +8,105 @@
 * `requests <https://requests.readthedocs.io/>`__
 """
 
+from __future__ import annotations
+
+import functools
 import logging
 import re
 from pathlib import Path
+from typing import TYPE_CHECKING
 
 from pyabc2._util import get_logger as _get_logger
 
+if TYPE_CHECKING:  # pragma: no cover
+    import requests
+
 logger = _get_logger(__name__)
 
 HERE = Path(__file__).parent
 
 SAVE_TO = HERE / "_bill-black"
 TXT_FNS = [
-    "a-tunes-1.txt",
-    "b-tunes-1.txt",
-    "c-tunes-1.txt",
-    "d-tunes-1.txt",
-    "e-tunes-1.txt",
-    "f-tunes-1.txt",
-    "g-tunes-1.txt",
-    "h-tunes-1.txt",
-    "i-tunes-1.txt",
-    "j-tunes-1.txt",
-    "k-tunes-1.txt",
-    "l-tunes-1.txt",
-    "m-tunes-1.txt",
-    "n-tunes-1.txt",
-    "o-tunes-1.txt",
-    "pq-tunes-1.txt",
-    "r-tunes-1.txt",
-    "s-tunes-2.rtf",
-    "t-tunes-1.txt",
-    "uv-tunes-1.txt",
-    "wz-tunes-1.txt",
+    "a-tunes.txt",
+    "b-tunes.txt",
+    "c-tunes.txt",
+    "d-tunes.txt",
+    "e-tunes.txt",
+    "f-tunes.txt",
+    "g-tunes.txt",
+    "h-tunes.txt",
+    "i-tunes.txt",
+    "j-tunes.txt",
+    "k-tunes.txt",
+    "L-tunes.txt",
+    "m-tunes.txt",
+    "n-tunes.txt",
+    "o-tunes.txt",
+    "pq-tunes.txt",
+    "r-tunes.txt",
+    "s-tunes.txt",
+    "t-tunes.txt",
+    "uv-tunes.txt",
+    "wxyz-tunes.txt",
 ]
 
 
+@functools.lru_cache(1)
+def _get_session() -> requests.Session:
+    return _build_session()
+
+
+def _build_session() -> requests.Session:
+    import requests
+    from requests.adapters import HTTPAdapter
+    from urllib3.util import Retry
+
+    session = requests.Session()
+    session.headers.update({"User-Agent": "pyabc2"})
+    retries = Retry(
+        total=5,
+        backoff_factor=0.5,
+        backoff_jitter=0.5,
+        allowed_methods={"GET", "HEAD"},
+        status_forcelist=[403, 429, 500, 502, 503, 504],
+        # Bill Black seems to sporadically return 403 (forbidden)
+        # possibly to indicate a temporary server issue or throttling/anti-bot
+    )
+    session.mount("https://", HTTPAdapter(max_retries=retries))
+    session.mount("http://", HTTPAdapter(max_retries=retries))
+
+    return session
+
+
 def download() -> None:
-    """Download the alphabetical text files from https://www.capeirish.com/ittl/alltunes/text/
+    """Download the alphabetical text files from https://www.capeirish.com/ittl/alltunes/alltunes-text/
     and store them in a compressed archive.
     """
+    import threading
     import zipfile
     from concurrent.futures import ThreadPoolExecutor
 
-    import requests
+    thread_local = threading.local()
+
+    def get_worker_session() -> requests.Session:
+        # One Session per worker thread to avoid cross-thread Session sharing.
+        try:
+            return thread_local.session
+        except AttributeError:
+            session = _build_session()
+            thread_local.session = session
+            return session
 
     def download_one(url):
-        r = requests.get(url, headers={"User-Agent": "pyabc2"}, timeout=5)
+        session = get_worker_session()
+        r = session.get(url, timeout=5)
         r.raise_for_status()
         return r.text
 
     with ThreadPoolExecutor(max_workers=4) as executor:
         futures = []
         for fn in TXT_FNS:
-            url = f"https://www.capeirish.com/ittl/alltunes/text/{fn}"
+            url = f"https://www.capeirish.com/ittl/alltunes/alltunes-text/{fn}"
             futures.append(executor.submit(download_one, url))
 
     SAVE_TO.mkdir(exist_ok=True)

pyabc2/sources/bill_black_tunefolders.py

@@ -4,8 +4,8 @@
 https://www.capeirish.com/ittl/
 
 As of the 2025-06-14 update, the "tunefolders" method is deprecated.
-Bill Black is now using the Eskin ABC Tools (https://www.capeirish.com/ittl/alltunes/html/),
-while also posting ABC text files (https://www.capeirish.com/ittl/alltunes/text/),
+Bill Black is now using the Eskin ABC Tools (https://www.capeirish.com/ittl/alltunes/alltunes-html/),
+while also posting ABC text files (https://www.capeirish.com/ittl/alltunes/alltunes-text/),
 both split up alphabetically by tune name.
 
 Requires:
@@ -19,6 +19,7 @@
 from pathlib import Path
 
 from pyabc2._util import get_logger as _get_logger
+from pyabc2.sources.bill_black import _get_session
 
 logger = _get_logger(__name__)
 
@@ -209,7 +210,7 @@ def get_collection(key: str) -> Collection:
 def download(key: str | Iterable[str] | None = None) -> None:
     import gzip
 
-    import requests
+    session = _get_session()
 
     SAVE_TO.mkdir(exist_ok=True)
 
@@ -224,7 +225,7 @@ def download(key: str | Iterable[str] | None = None) -> None:
         for url in collection.abc_urls:
             p = collection.url_to_file(url)
             logger.info(f"Downloading {url} to {p.relative_to(HERE).as_posix()}")
-            r = requests.get(url, headers={"User-Agent": "pyabc2"}, timeout=5)
+            r = session.get(url, timeout=5)
             r.raise_for_status()
 
             # Extract filename from URL and append .gz