urun-sense 0.1.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- urun_sense-0.1.0/.gitignore +4 -0
- urun_sense-0.1.0/LICENSE +201 -0
- urun_sense-0.1.0/PKG-INFO +139 -0
- urun_sense-0.1.0/README.md +118 -0
- urun_sense-0.1.0/pyproject.toml +32 -0
- urun_sense-0.1.0/src/sense/__init__.py +170 -0
- urun_sense-0.1.0/src/sense/_core.py +952 -0
- urun_sense-0.1.0/src/sense/_session.py +804 -0
- urun_sense-0.1.0/src/sense/_transport.py +464 -0
- urun_sense-0.1.0/src/sense/examples/__init__.py +1 -0
- urun_sense-0.1.0/src/sense/examples/assets/mic-question-48k-mono.wav +0 -0
- urun_sense-0.1.0/src/sense/examples/cheese_demo.py +466 -0
- urun_sense-0.1.0/src/sense/examples/voice_agent.py +170 -0
- urun_sense-0.1.0/tests/test_cheese_demo.py +347 -0
- urun_sense-0.1.0/tests/test_examples.py +25 -0
- urun_sense-0.1.0/tests/test_sdk.py +810 -0
- urun_sense-0.1.0/tests/test_transport.py +350 -0
- urun_sense-0.1.0/uv.lock +1173 -0
urun_sense-0.1.0/LICENSE
ADDED
|
@@ -0,0 +1,201 @@
|
|
|
1
|
+
Apache License
|
|
2
|
+
Version 2.0, January 2004
|
|
3
|
+
http://www.apache.org/licenses/
|
|
4
|
+
|
|
5
|
+
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
|
|
6
|
+
|
|
7
|
+
1. Definitions.
|
|
8
|
+
|
|
9
|
+
"License" shall mean the terms and conditions for use, reproduction,
|
|
10
|
+
and distribution as defined by Sections 1 through 9 of this document.
|
|
11
|
+
|
|
12
|
+
"Licensor" shall mean the copyright owner or entity authorized by
|
|
13
|
+
the copyright owner that are granting the License.
|
|
14
|
+
|
|
15
|
+
"Legal Entity" shall mean the union of the acting entity and all
|
|
16
|
+
other entities that control, are controlled by, or are under common
|
|
17
|
+
control with that entity. For the purposes of this definition,
|
|
18
|
+
"control" means (i) the power, direct or indirect, to cause the
|
|
19
|
+
direction or management of such entity, whether by contract or
|
|
20
|
+
otherwise, or (ii) ownership of fifty percent (50%) or more of the
|
|
21
|
+
outstanding shares, or (iii) beneficial ownership of such entity.
|
|
22
|
+
|
|
23
|
+
"You" (or "Your") shall mean an individual or Legal Entity
|
|
24
|
+
exercising permissions granted by this License.
|
|
25
|
+
|
|
26
|
+
"Source" form shall mean the preferred form for making modifications,
|
|
27
|
+
including but not limited to software source code, documentation
|
|
28
|
+
source, and configuration files.
|
|
29
|
+
|
|
30
|
+
"Object" form shall mean any form resulting from mechanical
|
|
31
|
+
transformation or translation of a Source form, including but
|
|
32
|
+
not limited to compiled object code, generated documentation,
|
|
33
|
+
and conversions to other media types.
|
|
34
|
+
|
|
35
|
+
"Work" shall mean the work of authorship, whether in Source or
|
|
36
|
+
Object form, made available under the License, as indicated by a
|
|
37
|
+
copyright notice that is included in or attached to the work
|
|
38
|
+
(an example is provided in the Appendix below).
|
|
39
|
+
|
|
40
|
+
"Derivative Works" shall mean any work, whether in Source or Object
|
|
41
|
+
form, that is based on (or derived from) the Work and for which the
|
|
42
|
+
editorial revisions, annotations, elaborations, or other modifications
|
|
43
|
+
represent, as a whole, an original work of authorship. For the purposes
|
|
44
|
+
of this License, Derivative Works shall not include works that remain
|
|
45
|
+
separable from, or merely link (or bind by name) to the interfaces of,
|
|
46
|
+
the Work and Derivative Works thereof.
|
|
47
|
+
|
|
48
|
+
"Contribution" shall mean any work of authorship, including
|
|
49
|
+
the original version of the Work and any modifications or additions
|
|
50
|
+
to that Work or Derivative Works thereof, that is intentionally
|
|
51
|
+
submitted to Licensor for inclusion in the Work by the copyright owner
|
|
52
|
+
or by an individual or Legal Entity authorized to submit on behalf of
|
|
53
|
+
the copyright owner. For the purposes of this definition, "submitted"
|
|
54
|
+
means any form of electronic, verbal, or written communication sent
|
|
55
|
+
to the Licensor or its representatives, including but not limited to
|
|
56
|
+
communication on electronic mailing lists, source code control systems,
|
|
57
|
+
and issue tracking systems that are managed by, or on behalf of, the
|
|
58
|
+
Licensor for the purpose of discussing and improving the Work, but
|
|
59
|
+
excluding communication that is conspicuously marked or otherwise
|
|
60
|
+
designated in writing by the copyright owner as "Not a Contribution."
|
|
61
|
+
|
|
62
|
+
"Contributor" shall mean Licensor and any individual or Legal Entity
|
|
63
|
+
on behalf of whom a Contribution has been received by Licensor and
|
|
64
|
+
subsequently incorporated within the Work.
|
|
65
|
+
|
|
66
|
+
2. Grant of Copyright License. Subject to the terms and conditions of
|
|
67
|
+
this License, each Contributor hereby grants to You a perpetual,
|
|
68
|
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
|
69
|
+
copyright license to reproduce, prepare Derivative Works of,
|
|
70
|
+
publicly display, publicly perform, sublicense, and distribute the
|
|
71
|
+
Work and such Derivative Works in Source or Object form.
|
|
72
|
+
|
|
73
|
+
3. Grant of Patent License. Subject to the terms and conditions of
|
|
74
|
+
this License, each Contributor hereby grants to You a perpetual,
|
|
75
|
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
|
76
|
+
(except as stated in this section) patent license to make, have made,
|
|
77
|
+
use, offer to sell, sell, import, and otherwise transfer the Work,
|
|
78
|
+
where such license applies only to those patent claims licensable
|
|
79
|
+
by such Contributor that are necessarily infringed by their
|
|
80
|
+
Contribution(s) alone or by combination of their Contribution(s)
|
|
81
|
+
with the Work to which such Contribution(s) was submitted. If You
|
|
82
|
+
institute patent litigation against any entity (including a
|
|
83
|
+
cross-claim or counterclaim in a lawsuit) alleging that the Work
|
|
84
|
+
or a Contribution incorporated within the Work constitutes direct
|
|
85
|
+
or contributory patent infringement, then any patent licenses
|
|
86
|
+
granted to You under this License for that Work shall terminate
|
|
87
|
+
as of the date such litigation is filed.
|
|
88
|
+
|
|
89
|
+
4. Redistribution. You may reproduce and distribute copies of the
|
|
90
|
+
Work or Derivative Works thereof in any medium, with or without
|
|
91
|
+
modifications, and in Source or Object form, provided that You
|
|
92
|
+
meet the following conditions:
|
|
93
|
+
|
|
94
|
+
(a) You must give any other recipients of the Work or
|
|
95
|
+
Derivative Works a copy of this License; and
|
|
96
|
+
|
|
97
|
+
(b) You must cause any modified files to carry prominent notices
|
|
98
|
+
stating that You changed the files; and
|
|
99
|
+
|
|
100
|
+
(c) You must retain, in the Source form of any Derivative Works
|
|
101
|
+
that You distribute, all copyright, patent, trademark, and
|
|
102
|
+
attribution notices from the Source form of the Work,
|
|
103
|
+
excluding those notices that do not pertain to any part of
|
|
104
|
+
the Derivative Works; and
|
|
105
|
+
|
|
106
|
+
(d) If the Work includes a "NOTICE" text file as part of its
|
|
107
|
+
distribution, then any Derivative Works that You distribute must
|
|
108
|
+
include a readable copy of the attribution notices contained
|
|
109
|
+
within such NOTICE file, excluding those notices that do not
|
|
110
|
+
pertain to any part of the Derivative Works, in at least one
|
|
111
|
+
of the following places: within a NOTICE text file distributed
|
|
112
|
+
as part of the Derivative Works; within the Source form or
|
|
113
|
+
documentation, if provided along with the Derivative Works; or,
|
|
114
|
+
within a display generated by the Derivative Works, if and
|
|
115
|
+
wherever such third-party notices normally appear. The contents
|
|
116
|
+
of the NOTICE file are for informational purposes only and
|
|
117
|
+
do not modify the License. You may add Your own attribution
|
|
118
|
+
notices within Derivative Works that You distribute, alongside
|
|
119
|
+
or as an addendum to the NOTICE text from the Work, provided
|
|
120
|
+
that such additional attribution notices cannot be construed
|
|
121
|
+
as modifying the License.
|
|
122
|
+
|
|
123
|
+
You may add Your own copyright statement to Your modifications and
|
|
124
|
+
may provide additional or different license terms and conditions
|
|
125
|
+
for use, reproduction, or distribution of Your modifications, or
|
|
126
|
+
for any such Derivative Works as a whole, provided Your use,
|
|
127
|
+
reproduction, and distribution of the Work otherwise complies with
|
|
128
|
+
the conditions stated in this License.
|
|
129
|
+
|
|
130
|
+
5. Submission of Contributions. Unless You explicitly state otherwise,
|
|
131
|
+
any Contribution intentionally submitted for inclusion in the Work
|
|
132
|
+
by You to the Licensor shall be under the terms and conditions of
|
|
133
|
+
this License, without any additional terms or conditions.
|
|
134
|
+
Notwithstanding the above, nothing herein shall supersede or modify
|
|
135
|
+
the terms of any separate license agreement you may have executed
|
|
136
|
+
with Licensor regarding such Contributions.
|
|
137
|
+
|
|
138
|
+
6. Trademarks. This License does not grant permission to use the trade
|
|
139
|
+
names, trademarks, service marks, or product names of the Licensor,
|
|
140
|
+
except as required for reasonable and customary use in describing the
|
|
141
|
+
origin of the Work and reproducing the content of the NOTICE file.
|
|
142
|
+
|
|
143
|
+
7. Disclaimer of Warranty. Unless required by applicable law or
|
|
144
|
+
agreed to in writing, Licensor provides the Work (and each
|
|
145
|
+
Contributor provides its Contributions) on an "AS IS" BASIS,
|
|
146
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
|
|
147
|
+
implied, including, without limitation, any warranties or conditions
|
|
148
|
+
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
|
|
149
|
+
PARTICULAR PURPOSE. You are solely responsible for determining the
|
|
150
|
+
appropriateness of using or redistributing the Work and assume any
|
|
151
|
+
risks associated with Your exercise of permissions under this License.
|
|
152
|
+
|
|
153
|
+
8. Limitation of Liability. In no event and under no legal theory,
|
|
154
|
+
whether in tort (including negligence), contract, or otherwise,
|
|
155
|
+
unless required by applicable law (such as deliberate and grossly
|
|
156
|
+
negligent acts) or agreed to in writing, shall any Contributor be
|
|
157
|
+
liable to You for damages, including any direct, indirect, special,
|
|
158
|
+
incidental, or consequential damages of any character arising as a
|
|
159
|
+
result of this License or out of the use or inability to use the
|
|
160
|
+
Work (including but not limited to damages for loss of goodwill,
|
|
161
|
+
work stoppage, computer failure or malfunction, or any and all
|
|
162
|
+
other commercial damages or losses), even if such Contributor
|
|
163
|
+
has been advised of the possibility of such damages.
|
|
164
|
+
|
|
165
|
+
9. Accepting Warranty or Additional Liability. While redistributing
|
|
166
|
+
the Work or Derivative Works thereof, You may choose to offer,
|
|
167
|
+
and charge a fee for, acceptance of support, warranty, indemnity,
|
|
168
|
+
or other liability obligations and/or rights consistent with this
|
|
169
|
+
License. However, in accepting such obligations, You may act only
|
|
170
|
+
on Your own behalf and on Your sole responsibility, not on behalf
|
|
171
|
+
of any other Contributor, and only if You agree to indemnify,
|
|
172
|
+
defend, and hold each Contributor harmless for any liability
|
|
173
|
+
incurred by, or claims asserted against, such Contributor by reason
|
|
174
|
+
of your accepting any such warranty or additional liability.
|
|
175
|
+
|
|
176
|
+
END OF TERMS AND CONDITIONS
|
|
177
|
+
|
|
178
|
+
APPENDIX: How to apply the Apache License to your work.
|
|
179
|
+
|
|
180
|
+
To apply the Apache License to your work, attach the following
|
|
181
|
+
boilerplate notice, with the fields enclosed by brackets "[]"
|
|
182
|
+
replaced with your own identifying information. (Don't include
|
|
183
|
+
the brackets!) The text should be enclosed in the appropriate
|
|
184
|
+
comment syntax for the file format. We also recommend that a
|
|
185
|
+
file or class name and description of purpose be included on the
|
|
186
|
+
same "printed page" as the copyright notice for easier
|
|
187
|
+
identification within third-party archives.
|
|
188
|
+
|
|
189
|
+
Copyright 2026 Urun Platform, Inc
|
|
190
|
+
|
|
191
|
+
Licensed under the Apache License, Version 2.0 (the "License");
|
|
192
|
+
you may not use this file except in compliance with the License.
|
|
193
|
+
You may obtain a copy of the License at
|
|
194
|
+
|
|
195
|
+
http://www.apache.org/licenses/LICENSE-2.0
|
|
196
|
+
|
|
197
|
+
Unless required by applicable law or agreed to in writing, software
|
|
198
|
+
distributed under the License is distributed on an "AS IS" BASIS,
|
|
199
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
200
|
+
See the License for the specific language governing permissions and
|
|
201
|
+
limitations under the License.
|
|
@@ -0,0 +1,139 @@
|
|
|
1
|
+
Metadata-Version: 2.5
|
|
2
|
+
Name: urun-sense
|
|
3
|
+
Version: 0.1.0
|
|
4
|
+
Summary: Python SDK for the Sense realtime decision protocol
|
|
5
|
+
Project-URL: Homepage, https://github.com/urun-sh/sense-sdk
|
|
6
|
+
Project-URL: Repository, https://github.com/urun-sh/sense-sdk
|
|
7
|
+
Author: Urun Platform Inc
|
|
8
|
+
License-Expression: Apache-2.0
|
|
9
|
+
License-File: LICENSE
|
|
10
|
+
Requires-Python: >=3.10
|
|
11
|
+
Requires-Dist: aiohttp>=3.9
|
|
12
|
+
Provides-Extra: demo
|
|
13
|
+
Requires-Dist: rich>=13; extra == 'demo'
|
|
14
|
+
Requires-Dist: sounddevice>=0.5; (sys_platform != 'linux') and extra == 'demo'
|
|
15
|
+
Provides-Extra: quic
|
|
16
|
+
Requires-Dist: aioquic>=1.2; extra == 'quic'
|
|
17
|
+
Requires-Dist: msgpack>=1.0; extra == 'quic'
|
|
18
|
+
Provides-Extra: webrtc
|
|
19
|
+
Requires-Dist: aiortc>=1.9; extra == 'webrtc'
|
|
20
|
+
Description-Content-Type: text/markdown
|
|
21
|
+
|
|
22
|
+
# urun-sense (Python SDK)
|
|
23
|
+
|
|
24
|
+
Python SDK for the Sense realtime decision protocol (`docs/protocol.md`). Dependency:
|
|
25
|
+
`aiohttp` only.
|
|
26
|
+
|
|
27
|
+
```bash
|
|
28
|
+
uv add urun-sense # or: pip install urun-sense
|
|
29
|
+
```
|
|
30
|
+
|
|
31
|
+
## Quick start
|
|
32
|
+
|
|
33
|
+
```python
|
|
34
|
+
import asyncio
|
|
35
|
+
from sense import Sense
|
|
36
|
+
|
|
37
|
+
app = Sense(api_key="...") # or SENSE_API_KEY env
|
|
38
|
+
finished = app.noul("Did the speaker just finish their turn?", id="finished")
|
|
39
|
+
intent = app.choice("What does the user want?", ["support", "sales", "other"], id="intent")
|
|
40
|
+
urgency = app.score("How urgent is this?", ["low", "medium", "high"], id="urgency")
|
|
41
|
+
|
|
42
|
+
s = app.session()
|
|
43
|
+
|
|
44
|
+
|
|
45
|
+
async def main():
|
|
46
|
+
async for wave in s.run(audio=mic_stream): # async iterator / callable / iterable
|
|
47
|
+
print(
|
|
48
|
+
wave.wave_id,
|
|
49
|
+
wave.t_ms,
|
|
50
|
+
wave.result("finished").value, # bool
|
|
51
|
+
wave.result("intent").top, # argmax option
|
|
52
|
+
wave.result("urgency").p_at_least("medium"),
|
|
53
|
+
)
|
|
54
|
+
```
|
|
55
|
+
|
|
56
|
+
## Flagship demo
|
|
57
|
+
|
|
58
|
+
The wheel ships the cheese classifier example (`sense/examples/cheese_demo.py`): it streams your
|
|
59
|
+
microphone, classifies what you say, and renders a live ANSI display. Requires Python 3.11+.
|
|
60
|
+
|
|
61
|
+
```bash
|
|
62
|
+
pip install "urun-sense[demo]" # adds rich + sounddevice (non-Linux)
|
|
63
|
+
SENSE_API_KEY=<key> python -m sense.examples.cheese_demo
|
|
64
|
+
```
|
|
65
|
+
|
|
66
|
+
PowerShell: `$env:SENSE_API_KEY = "<key>"` first, then run the `python` line bare.
|
|
67
|
+
|
|
68
|
+
Linux drives PipeWire directly (`pw-record`/`pw-dump`/`wpctl`); macOS and Windows need microphone
|
|
69
|
+
permission for your terminal. `--mic-only` runs without a key.
|
|
70
|
+
|
|
71
|
+
## Surface
|
|
72
|
+
|
|
73
|
+
- **Questions** — TEXT-FIRST: `app.noul(question, *, id=None, options=None)`,
|
|
74
|
+
`app.choice(question, options, *, id=None)`, `app.score(question, levels, *, id=None)`.
|
|
75
|
+
The natural-language question is the first argument; the id is optional and
|
|
76
|
+
auto-slugged (lowercase, non-alnum → `_`, unique) when omitted. Low-level
|
|
77
|
+
alternative: `Session.add_question(qid, kind, ...)`.
|
|
78
|
+
- **Predicates** — questions and accessors compose into `Predicate` trees:
|
|
79
|
+
`finished & (intent == "support")`, `finished.p >= 0.8`,
|
|
80
|
+
`intent.p("support") >= 0.7`, `urgency.value >= 1.5`,
|
|
81
|
+
`urgency.p_at_least("medium") >= 0.6`, `~q`, `Predicate.all_of/any_of`,
|
|
82
|
+
`Predicate.between(subject, enter, exit)` (stateful hysteresis latch).
|
|
83
|
+
Score comparison rule: on the bare question subject, `q == x` / `q != x`
|
|
84
|
+
compare the TOP LEVEL NAME (argmax) while `q > x` / `>=` / `<` / `<=`
|
|
85
|
+
compare the EXPECTED ORDINAL.
|
|
86
|
+
`q == x` and `q.top == x` are the same argmax comparison on a choice question.
|
|
87
|
+
- **Context** — passed to `when`/`onWave` handlers (and usable standalone for
|
|
88
|
+
offline replay): `ctx.latest(source=None)`, `ctx.waves(coalesce=False)`
|
|
89
|
+
(bounded to the last 512 waves), `ctx.when(pred, handler, source=None) -> handle`
|
|
90
|
+
(edge-triggered), `ctx.onWave(handler, source=None)`, `ctx.get(q, source=None)`,
|
|
91
|
+
`ctx.run(key, fn(ctx), on_cancel=..., replace=False)`, `ctx.cancel(key, barge_in=True)`,
|
|
92
|
+
`ctx.signal` (the current run's cancellation event, TS AbortSignal parity),
|
|
93
|
+
`ctx.session`.
|
|
94
|
+
- **Inputs** — `s.audio.push(pcm_bytes, source="default")` (PCM16 24 kHz mono),
|
|
95
|
+
`s.video.push(jpeg_bytes)` (≤ 256 KB/frame), `s.text.push(text, mode="append"|"replace"|"event")`.
|
|
96
|
+
A text push alone opens and drives a turn — a text-only session works with
|
|
97
|
+
zero audio.
|
|
98
|
+
- **Session** — `s.latest(q, source=None)`, `s.when(..., source=None)`,
|
|
99
|
+
`s.onWave(handler, coalesce=False, source=None)`, `s.waves(coalesce=False)`
|
|
100
|
+
(live bounded subscription — see below), `s.stats`
|
|
101
|
+
(waves/decisions/errors/protocol_errors), `s.close()` (cancels every tracked
|
|
102
|
+
ctx task, firing `on_cancel` hooks, before tearing down the wire).
|
|
103
|
+
|
|
104
|
+
## Wire behavior (protocol.md §1–§7)
|
|
105
|
+
|
|
106
|
+
- `wss://.../v1/realtime?model=...` + Bearer auth; `session.update` carries the
|
|
107
|
+
question envelope (`{questions, hz}`) as a JSON string.
|
|
108
|
+
- One commit + `response.create(max_output_tokens=250_000)` opens a cadence turn;
|
|
109
|
+
the empty commit before audio arrives is benign and ignored (`empty_buffer`).
|
|
110
|
+
- Decision lines are assembled into waves client-side (§6.3: repeat-qid flush,
|
|
111
|
+
turn-end flush, `sense.wave` boundary preferred when PR-3 lands).
|
|
112
|
+
- An incomplete wave (missing/unknown/null answer) is a `ProtocolError`: logged
|
|
113
|
+
loudly and counted in `stats.protocol_errors` — never silently treated as
|
|
114
|
+
false/hold.
|
|
115
|
+
- `503`/`429` back off (5 s→30 s cap, `Retry-After` honored), `401`/`404` fatal,
|
|
116
|
+
`session_gone` (close 1011) reconnects fresh, `buffer_full` commits.
|
|
117
|
+
- `ctx.cancel(key)` / barge-in sends `response.cancel`; `ctx.cancel(key, barge_in=False)`
|
|
118
|
+
cancels background work without preempting the turn (used by
|
|
119
|
+
`run(..., replace=True)` supersede).
|
|
120
|
+
|
|
121
|
+
## Documented degradations (awaiting serving PRs)
|
|
122
|
+
|
|
123
|
+
- **Audio sources (PR-2)**: non-default sources are queued client-side and pushed
|
|
124
|
+
round-robin over the single unlabeled audio lane; every wave is labeled with
|
|
125
|
+
the input source it was drained for, and `when`/`onWave`/`latest` accept
|
|
126
|
+
`source=` to evaluate against that source's stream. Sources are a client-side
|
|
127
|
+
concept until the server PR lands — no `source` field is sent on the wire.
|
|
128
|
+
- **Session.waves() vs run()**: `run()` is the primary consumption path;
|
|
129
|
+
`s.waves()` is a live bounded subscription (queue of 64, drop-oldest) that
|
|
130
|
+
never replays history and never competes with `run()` delivery. `ctx.waves()`
|
|
131
|
+
keeps the last 512 waves for offline replay.
|
|
132
|
+
- **Text replace/event**: the conversation transcript is append-only on today's
|
|
133
|
+
wire; `replace`/`event` are client-side bookkeeping with documented degradation.
|
|
134
|
+
|
|
135
|
+
## Tests
|
|
136
|
+
|
|
137
|
+
```bash
|
|
138
|
+
uv run --project . pytest tests ../conformance/python_runner.py
|
|
139
|
+
```
|
|
@@ -0,0 +1,118 @@
|
|
|
1
|
+
# urun-sense (Python SDK)
|
|
2
|
+
|
|
3
|
+
Python SDK for the Sense realtime decision protocol (`docs/protocol.md`). Dependency:
|
|
4
|
+
`aiohttp` only.
|
|
5
|
+
|
|
6
|
+
```bash
|
|
7
|
+
uv add urun-sense # or: pip install urun-sense
|
|
8
|
+
```
|
|
9
|
+
|
|
10
|
+
## Quick start
|
|
11
|
+
|
|
12
|
+
```python
|
|
13
|
+
import asyncio
|
|
14
|
+
from sense import Sense
|
|
15
|
+
|
|
16
|
+
app = Sense(api_key="...") # or SENSE_API_KEY env
|
|
17
|
+
finished = app.noul("Did the speaker just finish their turn?", id="finished")
|
|
18
|
+
intent = app.choice("What does the user want?", ["support", "sales", "other"], id="intent")
|
|
19
|
+
urgency = app.score("How urgent is this?", ["low", "medium", "high"], id="urgency")
|
|
20
|
+
|
|
21
|
+
s = app.session()
|
|
22
|
+
|
|
23
|
+
|
|
24
|
+
async def main():
|
|
25
|
+
async for wave in s.run(audio=mic_stream): # async iterator / callable / iterable
|
|
26
|
+
print(
|
|
27
|
+
wave.wave_id,
|
|
28
|
+
wave.t_ms,
|
|
29
|
+
wave.result("finished").value, # bool
|
|
30
|
+
wave.result("intent").top, # argmax option
|
|
31
|
+
wave.result("urgency").p_at_least("medium"),
|
|
32
|
+
)
|
|
33
|
+
```
|
|
34
|
+
|
|
35
|
+
## Flagship demo
|
|
36
|
+
|
|
37
|
+
The wheel ships the cheese classifier example (`sense/examples/cheese_demo.py`): it streams your
|
|
38
|
+
microphone, classifies what you say, and renders a live ANSI display. Requires Python 3.11+.
|
|
39
|
+
|
|
40
|
+
```bash
|
|
41
|
+
pip install "urun-sense[demo]" # adds rich + sounddevice (non-Linux)
|
|
42
|
+
SENSE_API_KEY=<key> python -m sense.examples.cheese_demo
|
|
43
|
+
```
|
|
44
|
+
|
|
45
|
+
PowerShell: `$env:SENSE_API_KEY = "<key>"` first, then run the `python` line bare.
|
|
46
|
+
|
|
47
|
+
Linux drives PipeWire directly (`pw-record`/`pw-dump`/`wpctl`); macOS and Windows need microphone
|
|
48
|
+
permission for your terminal. `--mic-only` runs without a key.
|
|
49
|
+
|
|
50
|
+
## Surface
|
|
51
|
+
|
|
52
|
+
- **Questions** — TEXT-FIRST: `app.noul(question, *, id=None, options=None)`,
|
|
53
|
+
`app.choice(question, options, *, id=None)`, `app.score(question, levels, *, id=None)`.
|
|
54
|
+
The natural-language question is the first argument; the id is optional and
|
|
55
|
+
auto-slugged (lowercase, non-alnum → `_`, unique) when omitted. Low-level
|
|
56
|
+
alternative: `Session.add_question(qid, kind, ...)`.
|
|
57
|
+
- **Predicates** — questions and accessors compose into `Predicate` trees:
|
|
58
|
+
`finished & (intent == "support")`, `finished.p >= 0.8`,
|
|
59
|
+
`intent.p("support") >= 0.7`, `urgency.value >= 1.5`,
|
|
60
|
+
`urgency.p_at_least("medium") >= 0.6`, `~q`, `Predicate.all_of/any_of`,
|
|
61
|
+
`Predicate.between(subject, enter, exit)` (stateful hysteresis latch).
|
|
62
|
+
Score comparison rule: on the bare question subject, `q == x` / `q != x`
|
|
63
|
+
compare the TOP LEVEL NAME (argmax) while `q > x` / `>=` / `<` / `<=`
|
|
64
|
+
compare the EXPECTED ORDINAL.
|
|
65
|
+
`q == x` and `q.top == x` are the same argmax comparison on a choice question.
|
|
66
|
+
- **Context** — passed to `when`/`onWave` handlers (and usable standalone for
|
|
67
|
+
offline replay): `ctx.latest(source=None)`, `ctx.waves(coalesce=False)`
|
|
68
|
+
(bounded to the last 512 waves), `ctx.when(pred, handler, source=None) -> handle`
|
|
69
|
+
(edge-triggered), `ctx.onWave(handler, source=None)`, `ctx.get(q, source=None)`,
|
|
70
|
+
`ctx.run(key, fn(ctx), on_cancel=..., replace=False)`, `ctx.cancel(key, barge_in=True)`,
|
|
71
|
+
`ctx.signal` (the current run's cancellation event, TS AbortSignal parity),
|
|
72
|
+
`ctx.session`.
|
|
73
|
+
- **Inputs** — `s.audio.push(pcm_bytes, source="default")` (PCM16 24 kHz mono),
|
|
74
|
+
`s.video.push(jpeg_bytes)` (≤ 256 KB/frame), `s.text.push(text, mode="append"|"replace"|"event")`.
|
|
75
|
+
A text push alone opens and drives a turn — a text-only session works with
|
|
76
|
+
zero audio.
|
|
77
|
+
- **Session** — `s.latest(q, source=None)`, `s.when(..., source=None)`,
|
|
78
|
+
`s.onWave(handler, coalesce=False, source=None)`, `s.waves(coalesce=False)`
|
|
79
|
+
(live bounded subscription — see below), `s.stats`
|
|
80
|
+
(waves/decisions/errors/protocol_errors), `s.close()` (cancels every tracked
|
|
81
|
+
ctx task, firing `on_cancel` hooks, before tearing down the wire).
|
|
82
|
+
|
|
83
|
+
## Wire behavior (protocol.md §1–§7)
|
|
84
|
+
|
|
85
|
+
- `wss://.../v1/realtime?model=...` + Bearer auth; `session.update` carries the
|
|
86
|
+
question envelope (`{questions, hz}`) as a JSON string.
|
|
87
|
+
- One commit + `response.create(max_output_tokens=250_000)` opens a cadence turn;
|
|
88
|
+
the empty commit before audio arrives is benign and ignored (`empty_buffer`).
|
|
89
|
+
- Decision lines are assembled into waves client-side (§6.3: repeat-qid flush,
|
|
90
|
+
turn-end flush, `sense.wave` boundary preferred when PR-3 lands).
|
|
91
|
+
- An incomplete wave (missing/unknown/null answer) is a `ProtocolError`: logged
|
|
92
|
+
loudly and counted in `stats.protocol_errors` — never silently treated as
|
|
93
|
+
false/hold.
|
|
94
|
+
- `503`/`429` back off (5 s→30 s cap, `Retry-After` honored), `401`/`404` fatal,
|
|
95
|
+
`session_gone` (close 1011) reconnects fresh, `buffer_full` commits.
|
|
96
|
+
- `ctx.cancel(key)` / barge-in sends `response.cancel`; `ctx.cancel(key, barge_in=False)`
|
|
97
|
+
cancels background work without preempting the turn (used by
|
|
98
|
+
`run(..., replace=True)` supersede).
|
|
99
|
+
|
|
100
|
+
## Documented degradations (awaiting serving PRs)
|
|
101
|
+
|
|
102
|
+
- **Audio sources (PR-2)**: non-default sources are queued client-side and pushed
|
|
103
|
+
round-robin over the single unlabeled audio lane; every wave is labeled with
|
|
104
|
+
the input source it was drained for, and `when`/`onWave`/`latest` accept
|
|
105
|
+
`source=` to evaluate against that source's stream. Sources are a client-side
|
|
106
|
+
concept until the server PR lands — no `source` field is sent on the wire.
|
|
107
|
+
- **Session.waves() vs run()**: `run()` is the primary consumption path;
|
|
108
|
+
`s.waves()` is a live bounded subscription (queue of 64, drop-oldest) that
|
|
109
|
+
never replays history and never competes with `run()` delivery. `ctx.waves()`
|
|
110
|
+
keeps the last 512 waves for offline replay.
|
|
111
|
+
- **Text replace/event**: the conversation transcript is append-only on today's
|
|
112
|
+
wire; `replace`/`event` are client-side bookkeeping with documented degradation.
|
|
113
|
+
|
|
114
|
+
## Tests
|
|
115
|
+
|
|
116
|
+
```bash
|
|
117
|
+
uv run --project . pytest tests ../conformance/python_runner.py
|
|
118
|
+
```
|
|
@@ -0,0 +1,32 @@
|
|
|
1
|
+
[project]
|
|
2
|
+
name = "urun-sense"
|
|
3
|
+
version = "0.1.0"
|
|
4
|
+
description = "Python SDK for the Sense realtime decision protocol"
|
|
5
|
+
readme = "README.md"
|
|
6
|
+
license = "Apache-2.0"
|
|
7
|
+
license-files = ["LICENSE"]
|
|
8
|
+
authors = [{ name = "Urun Platform Inc" }]
|
|
9
|
+
requires-python = ">=3.10"
|
|
10
|
+
dependencies = ["aiohttp>=3.9"]
|
|
11
|
+
|
|
12
|
+
[project.optional-dependencies]
|
|
13
|
+
# Transport extras: the core wire is WebSocket; QUIC (WebTransport) and
|
|
14
|
+
# WebRTC are opt-in and feature-detected at dial time.
|
|
15
|
+
quic = ["aioquic>=1.2", "msgpack>=1.0"]
|
|
16
|
+
webrtc = ["aiortc>=1.9"]
|
|
17
|
+
# Everything the shipped cheese demo needs beyond the SDK itself.
|
|
18
|
+
demo = ["rich>=13", 'sounddevice>=0.5; sys_platform != "linux"']
|
|
19
|
+
|
|
20
|
+
[project.urls]
|
|
21
|
+
Homepage = "https://github.com/urun-sh/sense-sdk"
|
|
22
|
+
Repository = "https://github.com/urun-sh/sense-sdk"
|
|
23
|
+
|
|
24
|
+
[build-system]
|
|
25
|
+
requires = ["hatchling>=1.26"]
|
|
26
|
+
build-backend = "hatchling.build"
|
|
27
|
+
|
|
28
|
+
[tool.hatch.build.targets.wheel]
|
|
29
|
+
packages = ["src/sense"]
|
|
30
|
+
|
|
31
|
+
[dependency-groups]
|
|
32
|
+
dev = ["pytest>=8", "rich>=13"]
|
|
@@ -0,0 +1,170 @@
|
|
|
1
|
+
"""sense — Python SDK for the Sense realtime decision protocol.
|
|
2
|
+
|
|
3
|
+
Public surface: :class:`Sense` (facade), :class:`Session`, :class:`Question`,
|
|
4
|
+
:class:`QuestionKind`, :class:`Predicate`, :class:`Wave`, :class:`Score`,
|
|
5
|
+
:class:`Result`, :class:`Context`, :class:`ProtocolError`,
|
|
6
|
+
:class:`SenseUnavailable`, :func:`session_event`, and the input handles
|
|
7
|
+
:class:`AudioInput`, :class:`VideoInput`, :class:`TextInput`. For clients that
|
|
8
|
+
drive their own event loop over a transport wire (qvs) the QUIC wire
|
|
9
|
+
:class:`QuicWire` and :class:`TransportFailed` are public too.
|
|
10
|
+
"""
|
|
11
|
+
|
|
12
|
+
from collections.abc import Callable
|
|
13
|
+
from typing import Any
|
|
14
|
+
|
|
15
|
+
from ._core import (
|
|
16
|
+
Context,
|
|
17
|
+
Predicate,
|
|
18
|
+
ProtocolError,
|
|
19
|
+
Question,
|
|
20
|
+
QuestionKind,
|
|
21
|
+
Result,
|
|
22
|
+
Score,
|
|
23
|
+
SessionEvent,
|
|
24
|
+
Wave,
|
|
25
|
+
session_event,
|
|
26
|
+
)
|
|
27
|
+
from ._session import (
|
|
28
|
+
AudioInput,
|
|
29
|
+
SenseUnavailable,
|
|
30
|
+
Session,
|
|
31
|
+
TextInput,
|
|
32
|
+
VideoInput,
|
|
33
|
+
)
|
|
34
|
+
from ._transport import QuicWire, TransportFailed
|
|
35
|
+
|
|
36
|
+
__all__ = [
|
|
37
|
+
"AudioInput",
|
|
38
|
+
"Context",
|
|
39
|
+
"Predicate",
|
|
40
|
+
"ProtocolError",
|
|
41
|
+
"Question",
|
|
42
|
+
"QuestionKind",
|
|
43
|
+
"QuicWire",
|
|
44
|
+
"Result",
|
|
45
|
+
"Score",
|
|
46
|
+
"Sense",
|
|
47
|
+
"SenseUnavailable",
|
|
48
|
+
"Session",
|
|
49
|
+
"SessionEvent",
|
|
50
|
+
"TextInput",
|
|
51
|
+
"TransportFailed",
|
|
52
|
+
"VideoInput",
|
|
53
|
+
"Wave",
|
|
54
|
+
"session_event",
|
|
55
|
+
]
|
|
56
|
+
|
|
57
|
+
|
|
58
|
+
def _slug(text: str, taken: set[str]) -> str:
|
|
59
|
+
"""Auto-id from the question text: lowercase, non-alnum -> ``_``, never
|
|
60
|
+
empty, unique within the app."""
|
|
61
|
+
base = "".join(c if c.isalnum() else "_" for c in text.strip().lower()).strip("_")
|
|
62
|
+
base = base or "question"
|
|
63
|
+
qid, n = base, 1
|
|
64
|
+
while qid in taken:
|
|
65
|
+
n += 1
|
|
66
|
+
qid = f"{base}_{n}"
|
|
67
|
+
return qid
|
|
68
|
+
|
|
69
|
+
|
|
70
|
+
class Sense:
|
|
71
|
+
"""Facade over a :class:`Session`: register questions, attach handlers,
|
|
72
|
+
then :meth:`session` for the wire loop.
|
|
73
|
+
|
|
74
|
+
Questions are TEXT-FIRST — the natural-language question is the first
|
|
75
|
+
positional argument; the id is optional and auto-slugged when omitted::
|
|
76
|
+
|
|
77
|
+
app = Sense(api_key=...)
|
|
78
|
+
finished = app.noul("Did the speaker just finish?")
|
|
79
|
+
intent = app.choice("What does the user want?", ["support", "sales", "other"])
|
|
80
|
+
urgency = app.score("How urgent is this?", ["low", "medium", "high"])
|
|
81
|
+
s = app.session()
|
|
82
|
+
s.when(finished & (intent == "support"), handle)
|
|
83
|
+
async for wave in s.run(audio=mic):
|
|
84
|
+
...
|
|
85
|
+
"""
|
|
86
|
+
|
|
87
|
+
def __init__(
|
|
88
|
+
self,
|
|
89
|
+
api_key: str | None = None,
|
|
90
|
+
base_url: str | None = None,
|
|
91
|
+
model: str = "sense-latest:bf16",
|
|
92
|
+
hz: float = 25.0,
|
|
93
|
+
transport: str = "auto",
|
|
94
|
+
) -> None:
|
|
95
|
+
self._kw = {"api_key": api_key, "base_url": base_url, "model": model, "hz": hz, "transport": transport}
|
|
96
|
+
self._session: Session | None = None
|
|
97
|
+
|
|
98
|
+
def _s(self) -> Session:
|
|
99
|
+
if self._session is None:
|
|
100
|
+
self._session = Session(**self._kw)
|
|
101
|
+
return self._session
|
|
102
|
+
|
|
103
|
+
def _qid(self, question: str, id: str | None) -> str:
|
|
104
|
+
if not isinstance(question, str) or not question.strip():
|
|
105
|
+
raise ValueError("question text must be a non-empty string (text-first API)")
|
|
106
|
+
if id is None:
|
|
107
|
+
return _slug(question, {q.id for q in self._s().questions})
|
|
108
|
+
if not id or not id.strip():
|
|
109
|
+
raise ValueError("explicit id must be non-empty")
|
|
110
|
+
return id
|
|
111
|
+
|
|
112
|
+
def noul(self, question: str, *, id: str | None = None, options: Any = None) -> Question:
|
|
113
|
+
return self._s().add_question(
|
|
114
|
+
self._qid(question, id), QuestionKind.NOUL, prompt=question, options=options
|
|
115
|
+
)
|
|
116
|
+
|
|
117
|
+
def choice(self, question: str, options: Any, *, id: str | None = None) -> Question:
|
|
118
|
+
return self._s().add_question(
|
|
119
|
+
self._qid(question, id), QuestionKind.CHOICE, prompt=question, options=options
|
|
120
|
+
)
|
|
121
|
+
|
|
122
|
+
def score(self, question: str, levels: Any, *, id: str | None = None) -> Question:
|
|
123
|
+
return self._s().add_question(
|
|
124
|
+
self._qid(question, id), QuestionKind.SCORE, prompt=question, levels=levels
|
|
125
|
+
)
|
|
126
|
+
|
|
127
|
+
def predicate(self, fn: Callable) -> Predicate:
|
|
128
|
+
"""Callable leaf: ``Predicate(fn)`` where ``fn(wave) -> bool``."""
|
|
129
|
+
return Predicate(fn)
|
|
130
|
+
|
|
131
|
+
def when(
|
|
132
|
+
self, pred: Predicate, handler: Callable | None = None, *, source: str | None = None
|
|
133
|
+
) -> Any:
|
|
134
|
+
"""Register an edge-triggered ``when``. Both forms work, and the
|
|
135
|
+
decorator form returns the function::
|
|
136
|
+
|
|
137
|
+
app.when(pred, handler)
|
|
138
|
+
@app.when(pred)
|
|
139
|
+
async def respond(event): ...
|
|
140
|
+
"""
|
|
141
|
+
if handler is None:
|
|
142
|
+
|
|
143
|
+
def deco(fn: Callable) -> Any:
|
|
144
|
+
self._s().when(pred, fn, source=source)
|
|
145
|
+
return fn
|
|
146
|
+
|
|
147
|
+
return deco
|
|
148
|
+
return self._s().when(pred, handler, source=source)
|
|
149
|
+
|
|
150
|
+
def onWave(
|
|
151
|
+
self, handler: Callable | None = None, coalesce: bool = False, *, source: str | None = None
|
|
152
|
+
) -> Any:
|
|
153
|
+
"""Observe every completed wave. Both forms work::
|
|
154
|
+
|
|
155
|
+
app.onWave(handler, coalesce=True)
|
|
156
|
+
@app.onWave(coalesce=True)
|
|
157
|
+
def observe(event): ...
|
|
158
|
+
"""
|
|
159
|
+
if handler is None:
|
|
160
|
+
|
|
161
|
+
def deco(fn: Callable) -> Any:
|
|
162
|
+
self._s().onWave(fn, coalesce, source=source)
|
|
163
|
+
return fn
|
|
164
|
+
|
|
165
|
+
return deco
|
|
166
|
+
self._s().onWave(handler, coalesce, source=source)
|
|
167
|
+
return handler
|
|
168
|
+
|
|
169
|
+
def session(self) -> Session:
|
|
170
|
+
return self._s()
|