stackvox 0.12.0__tar.gz → 0.13.0__tar.gz

{stackvox-0.12.0/stackvox.egg-info → stackvox-0.13.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: stackvox
-Version: 0.12.0
+Version: 0.13.0
 Summary: Offline TTS library using Kokoro-82M
 Author-email: "StackOne Technologies Ltd." <support@stackone.com>
 License-Expression: Apache-2.0
@@ -103,10 +103,12 @@ stackvox speak --normalize --file post.md       # normalize, then synthesize
 stackvox say   --normalize "**Shipped** 1,198.9 MPG"   # normalize, then send to daemon
 ```
 
-`--normalize` is off by default on `speak`/`say`, so notification phrases synthesize exactly as before. The tuning flags mirror the library defaults (all on except emoji-strip): `--no-markdown`, `--no-dev-terms`, `--no-expand-units`, `--no-expand-numbers`, `--no-pauses`, `--tables {drop,csv}`, `--code-blocks {drop,placeholder}`, `--strip-emoji`, `--no-terminal-stops`, `--locale`, and a `--pronunciations FILE` (a JSON `{"written": "spoken"}` map, applied whole-word and case-insensitively).
+`--normalize` is off by default on `speak`/`say`, so notification phrases synthesize exactly as before. The tuning flags mirror the library defaults (all on except emoji-strip): `--no-markdown`, `--no-dev-terms`, `--no-abbreviations`, `--no-filenames`, `--no-expand-units`, `--no-expand-numbers`, `--no-pauses`, `--tables {drop,csv}`, `--code-blocks {drop,placeholder}`, `--strip-emoji`, `--no-terminal-stops`, `--locale`, and a `--pronunciations FILE` (a JSON `{"written": "spoken"}` map, applied whole-word and case-insensitively).
 
 **Dev-output handling** (on by default; `--no-dev-terms` to disable) covers two things espeak reads badly: acronyms it says as words (`CLI` → "C L I", `AWS`, `URI`, `IAM`, …) and file-line references, which it voices literally as "dot py colon". A `file.ext:line` ref is re-spoken lead-with-the-line: `engine.py:42` → "line 42 of engine py", `cli.py:100-118` → "lines 100 to 118 of cli py", `foo.ts:666:10` → "line 666, column 10 of foo ts". Times, ratios, and versions (`12:30`, `3:1`, `1.2.3`) are left alone.
 
+**Chat shorthand** (on by default; `--no-abbreviations` to disable) is expanded to the words it stands for, so a response that reads "lmk" is spoken as "let me know": `lmk`, `iirc`, `afaik`, `idk`, `tbh`, `imo`, `btw`, `fyi`, `asap`, `wrt`, `tl;dr`/`tldr`. Whole-word and case-insensitive. Shorthand that doubles as a real word or is already voiced correctly (`rn`, `eta`, `aka`) is deliberately left alone; a `--pronunciations` entry overrides any of these.
+
 Fenced code blocks are dropped by default. `--code-blocks placeholder` instead speaks a short stand-in so a skipped block doesn't sound disjointed (ChatGPT-style) — set the wording with `--code-placeholder` (e.g. `--code-blocks placeholder --code-placeholder "you can see the code in the chat history"`); consecutive blocks collapse to one.
 
 Bash completion:

{stackvox-0.12.0 → stackvox-0.13.0}/README.md

@@ -69,10 +69,12 @@ stackvox speak --normalize --file post.md       # normalize, then synthesize
 stackvox say   --normalize "**Shipped** 1,198.9 MPG"   # normalize, then send to daemon
 ```
 
-`--normalize` is off by default on `speak`/`say`, so notification phrases synthesize exactly as before. The tuning flags mirror the library defaults (all on except emoji-strip): `--no-markdown`, `--no-dev-terms`, `--no-expand-units`, `--no-expand-numbers`, `--no-pauses`, `--tables {drop,csv}`, `--code-blocks {drop,placeholder}`, `--strip-emoji`, `--no-terminal-stops`, `--locale`, and a `--pronunciations FILE` (a JSON `{"written": "spoken"}` map, applied whole-word and case-insensitively).
+`--normalize` is off by default on `speak`/`say`, so notification phrases synthesize exactly as before. The tuning flags mirror the library defaults (all on except emoji-strip): `--no-markdown`, `--no-dev-terms`, `--no-abbreviations`, `--no-filenames`, `--no-expand-units`, `--no-expand-numbers`, `--no-pauses`, `--tables {drop,csv}`, `--code-blocks {drop,placeholder}`, `--strip-emoji`, `--no-terminal-stops`, `--locale`, and a `--pronunciations FILE` (a JSON `{"written": "spoken"}` map, applied whole-word and case-insensitively).
 
 **Dev-output handling** (on by default; `--no-dev-terms` to disable) covers two things espeak reads badly: acronyms it says as words (`CLI` → "C L I", `AWS`, `URI`, `IAM`, …) and file-line references, which it voices literally as "dot py colon". A `file.ext:line` ref is re-spoken lead-with-the-line: `engine.py:42` → "line 42 of engine py", `cli.py:100-118` → "lines 100 to 118 of cli py", `foo.ts:666:10` → "line 666, column 10 of foo ts". Times, ratios, and versions (`12:30`, `3:1`, `1.2.3`) are left alone.
 
+**Chat shorthand** (on by default; `--no-abbreviations` to disable) is expanded to the words it stands for, so a response that reads "lmk" is spoken as "let me know": `lmk`, `iirc`, `afaik`, `idk`, `tbh`, `imo`, `btw`, `fyi`, `asap`, `wrt`, `tl;dr`/`tldr`. Whole-word and case-insensitive. Shorthand that doubles as a real word or is already voiced correctly (`rn`, `eta`, `aka`) is deliberately left alone; a `--pronunciations` entry overrides any of these.
+
 Fenced code blocks are dropped by default. `--code-blocks placeholder` instead speaks a short stand-in so a skipped block doesn't sound disjointed (ChatGPT-style) — set the wording with `--code-placeholder` (e.g. `--code-blocks placeholder --code-placeholder "you can see the code in the chat history"`); consecutive blocks collapse to one.
 
 Bash completion:

{stackvox-0.12.0 → stackvox-0.13.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "stackvox"
-version = "0.12.0"
+version = "0.13.0"
 description = "Offline TTS library using Kokoro-82M"
 readme = "README.md"
 requires-python = ">=3.10"

{stackvox-0.12.0 → stackvox-0.13.0}/stackvox/cli.py

@@ -87,7 +87,7 @@ _stackvox_completion() {
             ;;
     esac
 
-    local norm_flags="--no-markdown --no-dev-terms --no-filenames --pronunciations --no-expand-units --no-expand-numbers --no-pauses --tables --code-blocks --code-placeholder --strip-emoji --no-terminal-stops --locale"
+    local norm_flags="--no-markdown --no-dev-terms --no-abbreviations --no-filenames --pronunciations --no-expand-units --no-expand-numbers --no-pauses --tables --code-blocks --code-placeholder --strip-emoji --no-terminal-stops --locale"
 
     case "$subcommand" in
         speak)
@@ -220,6 +220,12 @@ def _add_normalize_args(parser: argparse.ArgumentParser, *, with_switch: bool) -
         action="store_false",
         help="Do not spell out dev acronyms espeak mispronounces (CLI, CI, IDE, AWS, URI, IAM, ...)",
     )
+    parser.add_argument(
+        "--no-abbreviations",
+        dest="abbreviations",
+        action="store_false",
+        help="Leave chat shorthand as written (no 'lmk' -> 'let me know', 'iirc', 'btw', ...)",
+    )
     parser.add_argument(
         "--no-filenames",
         dest="filenames",
@@ -300,6 +306,7 @@ def _normalize_kwargs(args: argparse.Namespace) -> dict:
         "markdown": args.markdown,
         "pronunciations": _load_pronunciations(args.pronunciations),
         "dev_terms": args.dev_terms,
+        "abbreviations": args.abbreviations,
         "filenames": args.filenames,
         "expand_units": args.expand_units,
         "expand_numbers": args.expand_numbers,

{stackvox-0.12.0 → stackvox-0.13.0}/stackvox/text.py

@@ -484,6 +484,36 @@ _DEV_PRONUNCIATIONS: dict[str, str] = {
 }
 
 
+# --------------------------------------------------------------------------- #
+# Abbreviations                                                               #
+# --------------------------------------------------------------------------- #
+# Chat shorthand, expanded to the words it stands for. Distinct from
+# _DEV_PRONUNCIATIONS above, which fixes how a term *sounds*; this changes
+# *which words are said*, so it gets its own flag (`abbreviations`). That is the
+# lesson `filenames` taught when it was coupled to `dev_terms`.
+#
+# Applied whole-word and case-insensitive, so "LMK", "Lmk" and "lmk" all match.
+# Deliberately excluded ("rn", "eta", "aka" and friends): real words, Greek
+# letters, or acronyms espeak already voices correctly, so expanding them costs
+# more in false positives than it buys. "fyi" keeps its letters because that is how it is
+# said aloud, not "for your information".
+_ABBREVIATIONS: dict[str, str] = {
+    "lmk": "let me know",
+    "iirc": "if I recall correctly",
+    "afaik": "as far as I know",
+    "idk": "I don't know",
+    "tbh": "to be honest",
+    "imo": "in my opinion",
+    "btw": "by the way",
+    "fyi": "F Y I",
+    "asap": "as soon as possible",
+    "wrt": "with respect to",
+    # Both spellings of the same abbreviation; the bare form is at least as common.
+    "tl;dr": "too long, didn't read",
+    "tldr": "too long, didn't read",
+}
+
+
 # --------------------------------------------------------------------------- #
 # Pauses                                                                      #
 # --------------------------------------------------------------------------- #
@@ -687,6 +717,7 @@ def normalize_for_speech(
     markdown: bool = True,
     pronunciations: dict[str, str] | None = None,
     dev_terms: bool = True,
+    abbreviations: bool = True,
     filenames: bool = True,
     expand_units: bool = True,
     expand_numbers: bool = True,
@@ -701,9 +732,13 @@ def normalize_for_speech(
     """Normalize ``text`` into speakable prose. Returns paragraphs joined by
     newlines. See ``docs/speech-normalization.md`` for the full contract."""
     expand_units_flag, expand_numbers_flag = expand_units, expand_numbers
-    # Built-in dev-term fixes first; caller-supplied pronunciations override them
-    # (keyed case-insensitively, so a caller's "CLI" beats the default "cli").
-    effective_pronunciations: dict[str, str] = dict(_DEV_PRONUNCIATIONS) if dev_terms else {}
+    # Built-in fixes first; caller-supplied pronunciations override them (keyed
+    # case-insensitively, so a caller's "CLI" beats the default "cli"). The two
+    # built-in dicts share no keys, and neither one's spoken form contains a key
+    # of the other, so the sequential passes can't feed each other.
+    effective_pronunciations: dict[str, str] = dict(_ABBREVIATIONS) if abbreviations else {}
+    if dev_terms:
+        effective_pronunciations.update(_DEV_PRONUNCIATIONS)
     for written, spoken in (pronunciations or {}).items():
         effective_pronunciations[written.lower()] = spoken
 

{stackvox-0.12.0 → stackvox-0.13.0/stackvox.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: stackvox
-Version: 0.12.0
+Version: 0.13.0
 Summary: Offline TTS library using Kokoro-82M
 Author-email: "StackOne Technologies Ltd." <support@stackone.com>
 License-Expression: Apache-2.0
@@ -103,10 +103,12 @@ stackvox speak --normalize --file post.md       # normalize, then synthesize
 stackvox say   --normalize "**Shipped** 1,198.9 MPG"   # normalize, then send to daemon
 ```
 
-`--normalize` is off by default on `speak`/`say`, so notification phrases synthesize exactly as before. The tuning flags mirror the library defaults (all on except emoji-strip): `--no-markdown`, `--no-dev-terms`, `--no-expand-units`, `--no-expand-numbers`, `--no-pauses`, `--tables {drop,csv}`, `--code-blocks {drop,placeholder}`, `--strip-emoji`, `--no-terminal-stops`, `--locale`, and a `--pronunciations FILE` (a JSON `{"written": "spoken"}` map, applied whole-word and case-insensitively).
+`--normalize` is off by default on `speak`/`say`, so notification phrases synthesize exactly as before. The tuning flags mirror the library defaults (all on except emoji-strip): `--no-markdown`, `--no-dev-terms`, `--no-abbreviations`, `--no-filenames`, `--no-expand-units`, `--no-expand-numbers`, `--no-pauses`, `--tables {drop,csv}`, `--code-blocks {drop,placeholder}`, `--strip-emoji`, `--no-terminal-stops`, `--locale`, and a `--pronunciations FILE` (a JSON `{"written": "spoken"}` map, applied whole-word and case-insensitively).
 
 **Dev-output handling** (on by default; `--no-dev-terms` to disable) covers two things espeak reads badly: acronyms it says as words (`CLI` → "C L I", `AWS`, `URI`, `IAM`, …) and file-line references, which it voices literally as "dot py colon". A `file.ext:line` ref is re-spoken lead-with-the-line: `engine.py:42` → "line 42 of engine py", `cli.py:100-118` → "lines 100 to 118 of cli py", `foo.ts:666:10` → "line 666, column 10 of foo ts". Times, ratios, and versions (`12:30`, `3:1`, `1.2.3`) are left alone.
 
+**Chat shorthand** (on by default; `--no-abbreviations` to disable) is expanded to the words it stands for, so a response that reads "lmk" is spoken as "let me know": `lmk`, `iirc`, `afaik`, `idk`, `tbh`, `imo`, `btw`, `fyi`, `asap`, `wrt`, `tl;dr`/`tldr`. Whole-word and case-insensitive. Shorthand that doubles as a real word or is already voiced correctly (`rn`, `eta`, `aka`) is deliberately left alone; a `--pronunciations` entry overrides any of these.
+
 Fenced code blocks are dropped by default. `--code-blocks placeholder` instead speaks a short stand-in so a skipped block doesn't sound disjointed (ChatGPT-style) — set the wording with `--code-placeholder` (e.g. `--code-blocks placeholder --code-placeholder "you can see the code in the chat history"`); consecutive blocks collapse to one.
 
 Bash completion:

{stackvox-0.12.0 → stackvox-0.13.0}/tests/test_cli.py

@@ -129,6 +129,7 @@ def _norm_ns(text=None, file=None, **overrides):
         markdown=True,
         pronunciations=None,
         dev_terms=True,
+        abbreviations=True,
         filenames=True,
         expand_units=True,
         expand_numbers=True,

{stackvox-0.12.0 → stackvox-0.13.0}/tests/test_text.py

@@ -487,3 +487,44 @@ def test_caller_pronunciations_override_dev_terms():
     out = normalize_for_speech("Use the CLI.", markdown=False, pronunciations={"CLI": "command line"})
     assert "command line" in out
     assert "C L I" not in out
+
+
+def test_chat_abbreviations_are_expanded():
+    assert normalize_for_speech("lmk what you think", markdown=False) == "let me know what you think."
+    assert "if I recall correctly" in normalize_for_speech("iirc it shipped", markdown=False)
+    assert "by the way" in normalize_for_speech("btw the build is green", markdown=False)
+    assert "as soon as possible" in normalize_for_speech("ship it asap", markdown=False)
+    assert "with respect to" in normalize_for_speech("wrt the plan", markdown=False)
+
+
+def test_chat_abbreviations_are_case_insensitive():
+    assert "let me know" in normalize_for_speech("LMK if that works", markdown=False)
+    assert "to be honest" in normalize_for_speech("Tbh it is fine", markdown=False)
+
+
+def test_tldr_expands_in_both_spellings():
+    assert "too long, didn't read" in normalize_for_speech("tl;dr it works", markdown=False)
+    assert "too long, didn't read" in normalize_for_speech("tldr it works", markdown=False)
+
+
+def test_chat_abbreviations_are_whole_word_only():
+    # must not rewrite the middle of a longer token
+    assert "let me know" not in normalize_for_speech("lmkfoo", markdown=False)
+    assert "in my opinion" not in normalize_for_speech("imos", markdown=False)
+
+
+def test_chat_abbreviations_can_be_disabled():
+    assert normalize_for_speech("lmk soon", markdown=False, abbreviations=False) == "lmk soon."
+
+
+def test_caller_pronunciations_override_abbreviations():
+    out = normalize_for_speech("lmk soon", markdown=False, pronunciations={"LMK": "tell me"})
+    assert "tell me" in out
+    assert "let me know" not in out
+
+
+def test_abbreviations_independent_of_dev_terms():
+    # the `filenames` lesson: an unrelated switch must not silently disable this
+    out = normalize_for_speech("lmk about the CLI", markdown=False, dev_terms=False)
+    assert "let me know" in out
+    assert "C L I" not in out