zebracat 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.
- zebracat-0.1.0/.gitignore +7 -0
- zebracat-0.1.0/LICENSE +21 -0
- zebracat-0.1.0/PKG-INFO +202 -0
- zebracat-0.1.0/README.md +179 -0
- zebracat-0.1.0/openapi/openapi.json +17152 -0
- zebracat-0.1.0/pyproject.toml +44 -0
- zebracat-0.1.0/scripts/generate.py +394 -0
- zebracat-0.1.0/src/zebracat/__init__.py +46 -0
- zebracat-0.1.0/src/zebracat/_client.py +443 -0
- zebracat-0.1.0/src/zebracat/_errors.py +152 -0
- zebracat-0.1.0/src/zebracat/_generated/__init__.py +5 -0
- zebracat-0.1.0/src/zebracat/_generated/meta.py +110 -0
- zebracat-0.1.0/src/zebracat/_generated/operations.py +3481 -0
- zebracat-0.1.0/src/zebracat/_generated/types.py +1179 -0
- zebracat-0.1.0/src/zebracat/_types.py +57 -0
- zebracat-0.1.0/src/zebracat/_version.py +1 -0
- zebracat-0.1.0/src/zebracat/py.typed +0 -0
- zebracat-0.1.0/tests/conftest.py +43 -0
- zebracat-0.1.0/tests/stub_server.py +80 -0
- zebracat-0.1.0/tests/test_client.py +358 -0
zebracat-0.1.0/LICENSE
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
MIT License
|
|
2
|
+
|
|
3
|
+
Copyright (c) 2026 Zebracat
|
|
4
|
+
|
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
6
|
+
of this software and associated documentation files (the "Software"), to deal
|
|
7
|
+
in the Software without restriction, including without limitation the rights
|
|
8
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
9
|
+
copies of the Software, and to permit persons to whom the Software is
|
|
10
|
+
furnished to do so, subject to the following conditions:
|
|
11
|
+
|
|
12
|
+
The above copyright notice and this permission notice shall be included in all
|
|
13
|
+
copies or substantial portions of the Software.
|
|
14
|
+
|
|
15
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
16
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
17
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
18
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
19
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
20
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
21
|
+
SOFTWARE.
|
zebracat-0.1.0/PKG-INFO
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
1
|
+
Metadata-Version: 2.5
|
|
2
|
+
Name: zebracat
|
|
3
|
+
Version: 0.1.0
|
|
4
|
+
Summary: Official Python client for the Zebracat video API
|
|
5
|
+
Project-URL: Homepage, https://zebracat.ai
|
|
6
|
+
Project-URL: Documentation, https://api.zebracat.ai/docs
|
|
7
|
+
Project-URL: Source, https://github.com/zebracatai/zebracat-python
|
|
8
|
+
Author: Zebracat
|
|
9
|
+
License-Expression: MIT
|
|
10
|
+
License-File: LICENSE
|
|
11
|
+
Keywords: ai,api,sdk,video,zebracat
|
|
12
|
+
Classifier: Development Status :: 4 - Beta
|
|
13
|
+
Classifier: Intended Audience :: Developers
|
|
14
|
+
Classifier: Programming Language :: Python :: 3
|
|
15
|
+
Classifier: Programming Language :: Python :: 3 :: Only
|
|
16
|
+
Classifier: Topic :: Multimedia :: Video
|
|
17
|
+
Classifier: Typing :: Typed
|
|
18
|
+
Requires-Python: >=3.9
|
|
19
|
+
Requires-Dist: httpx<1,>=0.24
|
|
20
|
+
Provides-Extra: test
|
|
21
|
+
Requires-Dist: pytest>=7; extra == 'test'
|
|
22
|
+
Description-Content-Type: text/markdown
|
|
23
|
+
|
|
24
|
+
# Zebracat Python SDK
|
|
25
|
+
|
|
26
|
+
The official Python client for the [Zebracat API](https://api.zebracat.ai/docs): make videos with AI from a brief,
|
|
27
|
+
a script or a link, then read, quote, generate and download them.
|
|
28
|
+
|
|
29
|
+
```bash
|
|
30
|
+
pip install zebracat
|
|
31
|
+
```
|
|
32
|
+
|
|
33
|
+
Python 3.9 or newer. One dependency: `httpx`.
|
|
34
|
+
|
|
35
|
+
## Quickstart
|
|
36
|
+
|
|
37
|
+
Create an API key at <https://zebracat.ai/api> and export it:
|
|
38
|
+
|
|
39
|
+
```bash
|
|
40
|
+
export ZEBRACAT_API_KEY=sk-...
|
|
41
|
+
```
|
|
42
|
+
|
|
43
|
+
```python
|
|
44
|
+
from zebracat import Zebracat
|
|
45
|
+
|
|
46
|
+
client = Zebracat() # reads ZEBRACAT_API_KEY
|
|
47
|
+
|
|
48
|
+
# 1. Plan a video. Planning is free; nothing is charged until you generate.
|
|
49
|
+
video = client.create_video(
|
|
50
|
+
brief="How bees make honey",
|
|
51
|
+
auto_generate=False,
|
|
52
|
+
options={"seconds": 30, "aspect": "vertical", "media_type": "ai_video"},
|
|
53
|
+
)
|
|
54
|
+
|
|
55
|
+
# 2. Wait for the plan, then look at the price.
|
|
56
|
+
video = client.wait_for_video(video["task_id"]) # returns at plan_ready
|
|
57
|
+
print(client.quote_video(video["task_id"]))
|
|
58
|
+
|
|
59
|
+
# 3. Generate it (this spends credit) and wait for the file.
|
|
60
|
+
client.generate_video(video["task_id"])
|
|
61
|
+
video = client.wait_for_video(video["task_id"]) # returns at completed
|
|
62
|
+
print(video["video_url"])
|
|
63
|
+
```
|
|
64
|
+
|
|
65
|
+
Pass `auto_generate=True` to plan and generate in one call; `wait_for_video` then returns once the video is done.
|
|
66
|
+
|
|
67
|
+
## Every operation is a method
|
|
68
|
+
|
|
69
|
+
Each operation of the [OpenAPI document](openapi/openapi.json) is a method named after its `operationId`:
|
|
70
|
+
`create_video`, `get_video`, `list_videos`, `quote_video`, `generate_video`, `clone_video`, `translate_video`,
|
|
71
|
+
`list_voices`, `list_templates`, `upload_file`, `create_upload_link`, `create_webhook`, `list_capabilities`,
|
|
72
|
+
`get_account`, and the rest.
|
|
73
|
+
|
|
74
|
+
- Path parameters are positional: `client.get_video("6f1c…")`.
|
|
75
|
+
- Query parameters and body fields are keyword arguments: `client.create_video(brief="…", auto_generate=True)`.
|
|
76
|
+
- Leave an argument out to not send it; pass `None` to send JSON `null`. Fields the SDK does not know yet go in
|
|
77
|
+
`extra_body={...}`.
|
|
78
|
+
- Answers are plain `dict`s, typed as `TypedDict`s in `zebracat.types` for your editor.
|
|
79
|
+
- Files (`upload_file`, `create_clone_voice`, `upload_to_link`) take bytes, an open binary file, or
|
|
80
|
+
`(filename, bytes, content_type)`.
|
|
81
|
+
- Operations the API marks deprecated still work and raise a `DeprecationWarning`.
|
|
82
|
+
|
|
83
|
+
## Uploading a file
|
|
84
|
+
|
|
85
|
+
`upload_file` takes a picture in one call. For a video to translate, clone or swap, or a voice sample, use a one-time
|
|
86
|
+
upload link: `upload_with_link` asks for the link (`create_upload_link`) and sends the file to it.
|
|
87
|
+
|
|
88
|
+
```python
|
|
89
|
+
stored = client.upload_with_link("talk.mp4", use="translate") # a local path, bytes or an open file
|
|
90
|
+
video = client.translate_video(video_url=stored["url"], target_language="es")
|
|
91
|
+
```
|
|
92
|
+
|
|
93
|
+
`use` is `picture`, `translate`, `clone`, `swap` or `voice_clone`; it sets the accepted types and the size cap.
|
|
94
|
+
|
|
95
|
+
## Authentication
|
|
96
|
+
|
|
97
|
+
```python
|
|
98
|
+
Zebracat(api_key="sk-...") # X-API-KEY, billed from the API balance
|
|
99
|
+
|
|
100
|
+
Zebracat( # OAuth: Authorization: Bearer zoa_..., billed from plan credits
|
|
101
|
+
access_token="zoa_...",
|
|
102
|
+
refresh_token="zor_...",
|
|
103
|
+
client_id="your-oauth-client-id",
|
|
104
|
+
client_secret=None, # only for confidential clients
|
|
105
|
+
on_token_refresh=lambda tokens: save(tokens),
|
|
106
|
+
)
|
|
107
|
+
```
|
|
108
|
+
|
|
109
|
+
With a refresh token, a `401` makes the client swap it for a new pair once (`POST /api/oauth/token`, the API rotates
|
|
110
|
+
both tokens) and repeat the call. `on_token_refresh` receives the new tokens so you can store them.
|
|
111
|
+
|
|
112
|
+
`ZEBRACAT_BASE_URL` (or `base_url=`) points the client at another server.
|
|
113
|
+
|
|
114
|
+
## Retries
|
|
115
|
+
|
|
116
|
+
A call is repeated up to `max_retries` times (default 3) when the API answers `429`, `500`, `502`, `503` or `504`,
|
|
117
|
+
or the connection drops:
|
|
118
|
+
|
|
119
|
+
- A `Retry-After` header is honoured. A wait longer than `max_retry_wait` (default 60 s, e.g. a daily cap) is raised
|
|
120
|
+
as `RateLimitError` with `retry_after` instead of slept through.
|
|
121
|
+
- Without `Retry-After` the wait is 0.5 s, 1 s, 2 s… (at most 8 s) with jitter.
|
|
122
|
+
- A `429` is always retried: the API refused before doing anything.
|
|
123
|
+
- A `5xx` or a dropped connection is retried only when resending cannot do the work twice: reads, `PUT`, `DELETE`,
|
|
124
|
+
and the paid POSTs whose server replays the first answer for a repeated `Idempotency-Key` (`create_video`,
|
|
125
|
+
`clone_video`, `translate_video` and the v1 creators). Other POSTs raise at once, so a charge is never repeated.
|
|
126
|
+
|
|
127
|
+
Per call: `request_options={"max_retries": 0, "timeout": 120, "headers": {...}}`.
|
|
128
|
+
|
|
129
|
+
## Idempotency keys
|
|
130
|
+
|
|
131
|
+
Every POST that spends money (it can answer `402`, or it buys credit) sends an `Idempotency-Key`. The SDK makes one
|
|
132
|
+
per call and reuses it across that call's retries. Pass your own to make a retry from your side safe too:
|
|
133
|
+
|
|
134
|
+
```python
|
|
135
|
+
client.create_video(brief="…", auto_generate=True, idempotency_key=f"order-{order_id}")
|
|
136
|
+
```
|
|
137
|
+
|
|
138
|
+
The same key with a different body answers `422 idempotency_key_reused`.
|
|
139
|
+
|
|
140
|
+
## Waiting for a video
|
|
141
|
+
|
|
142
|
+
`wait_for_video(task_id, timeout=1800, poll_interval=None, on_update=None)` reads the video again after the
|
|
143
|
+
`poll_after_s` seconds the API gives (falling back to 10–30 s per phase on an API that does not send it). It returns
|
|
144
|
+
when nothing more happens on its own:
|
|
145
|
+
|
|
146
|
+
| phase | what you do next |
|
|
147
|
+
|---|---|
|
|
148
|
+
| `completed` | `video["video_url"]` is the file; `download_video_file(task_id)` returns the URL as well |
|
|
149
|
+
| `plan_ready` | show `quote_video(task_id)`, then `generate_video(task_id)` and wait again |
|
|
150
|
+
| `needs_input` | answer `video["question"]` with `answer_video_question(task_id, ...)` and wait again |
|
|
151
|
+
|
|
152
|
+
It raises `VideoFailedError` (`.video`, `.phase`, `.error`) when the video fails or is cancelled, and
|
|
153
|
+
`WaitTimeoutError` (`.video`) when `timeout` passes first; the video keeps going on our side.
|
|
154
|
+
|
|
155
|
+
## Errors
|
|
156
|
+
|
|
157
|
+
Every error from the API is a subclass of `zebracat.APIError` and carries the public error envelope:
|
|
158
|
+
|
|
159
|
+
```python
|
|
160
|
+
from zebracat import InsufficientFundsError, RateLimitError, APIError
|
|
161
|
+
|
|
162
|
+
try:
|
|
163
|
+
client.generate_video(task_id)
|
|
164
|
+
except InsufficientFundsError as e:
|
|
165
|
+
print("Add credit:", e.topup_url)
|
|
166
|
+
except RateLimitError as e:
|
|
167
|
+
print("Try again in", e.retry_after, "s")
|
|
168
|
+
except APIError as e:
|
|
169
|
+
print(e.status_code, e.code, e.hint, e.request_id)
|
|
170
|
+
```
|
|
171
|
+
|
|
172
|
+
| status | class |
|
|
173
|
+
|---|---|
|
|
174
|
+
| 400, 422 | `BadRequestError` (`field`, `allowed_values`, `errors`) |
|
|
175
|
+
| 401 | `AuthenticationError` |
|
|
176
|
+
| 402 | `InsufficientFundsError` (`topup_url`) |
|
|
177
|
+
| 403 | `PermissionDeniedError` (`upgrade_url`) |
|
|
178
|
+
| 404, 410 | `NotFoundError` |
|
|
179
|
+
| 409 | `ConflictError` |
|
|
180
|
+
| 429 | `RateLimitError` (`retry_after`) |
|
|
181
|
+
| 5xx | `InternalServerError` |
|
|
182
|
+
| no answer | `APIConnectionError`, `APITimeoutError` |
|
|
183
|
+
|
|
184
|
+
`request_id` is the id to quote to support: the API's `request_id` (also the `X-Request-ID` header). The SDK sends
|
|
185
|
+
an `X-Request-ID` of its own on every call, so even a connection error has one.
|
|
186
|
+
|
|
187
|
+
## Development
|
|
188
|
+
|
|
189
|
+
```bash
|
|
190
|
+
uv venv && uv pip install -e '.[test]'
|
|
191
|
+
python scripts/generate.py # regenerate src/zebracat/_generated from openapi/openapi.json
|
|
192
|
+
python scripts/generate.py --check # CI: fails when the generated code is out of date
|
|
193
|
+
pytest # runs against a local stub server, no network
|
|
194
|
+
```
|
|
195
|
+
|
|
196
|
+
`openapi/openapi.json` comes from the backend: `python manage.py spectacular --format openapi-json --file
|
|
197
|
+
openapi.json` (or download `https://api.zebracat.ai/docs/openapi.json`). Copy it here, regenerate, commit both.
|
|
198
|
+
|
|
199
|
+
## Releasing
|
|
200
|
+
|
|
201
|
+
Bump `src/zebracat/_version.py`, commit, tag `vX.Y.Z` and push the tag. `.github/workflows/release.yml` tests,
|
|
202
|
+
builds and publishes to PyPI with trusted publishing (GitHub OIDC; no token is stored anywhere).
|
zebracat-0.1.0/README.md
ADDED
|
@@ -0,0 +1,179 @@
|
|
|
1
|
+
# Zebracat Python SDK
|
|
2
|
+
|
|
3
|
+
The official Python client for the [Zebracat API](https://api.zebracat.ai/docs): make videos with AI from a brief,
|
|
4
|
+
a script or a link, then read, quote, generate and download them.
|
|
5
|
+
|
|
6
|
+
```bash
|
|
7
|
+
pip install zebracat
|
|
8
|
+
```
|
|
9
|
+
|
|
10
|
+
Python 3.9 or newer. One dependency: `httpx`.
|
|
11
|
+
|
|
12
|
+
## Quickstart
|
|
13
|
+
|
|
14
|
+
Create an API key at <https://zebracat.ai/api> and export it:
|
|
15
|
+
|
|
16
|
+
```bash
|
|
17
|
+
export ZEBRACAT_API_KEY=sk-...
|
|
18
|
+
```
|
|
19
|
+
|
|
20
|
+
```python
|
|
21
|
+
from zebracat import Zebracat
|
|
22
|
+
|
|
23
|
+
client = Zebracat() # reads ZEBRACAT_API_KEY
|
|
24
|
+
|
|
25
|
+
# 1. Plan a video. Planning is free; nothing is charged until you generate.
|
|
26
|
+
video = client.create_video(
|
|
27
|
+
brief="How bees make honey",
|
|
28
|
+
auto_generate=False,
|
|
29
|
+
options={"seconds": 30, "aspect": "vertical", "media_type": "ai_video"},
|
|
30
|
+
)
|
|
31
|
+
|
|
32
|
+
# 2. Wait for the plan, then look at the price.
|
|
33
|
+
video = client.wait_for_video(video["task_id"]) # returns at plan_ready
|
|
34
|
+
print(client.quote_video(video["task_id"]))
|
|
35
|
+
|
|
36
|
+
# 3. Generate it (this spends credit) and wait for the file.
|
|
37
|
+
client.generate_video(video["task_id"])
|
|
38
|
+
video = client.wait_for_video(video["task_id"]) # returns at completed
|
|
39
|
+
print(video["video_url"])
|
|
40
|
+
```
|
|
41
|
+
|
|
42
|
+
Pass `auto_generate=True` to plan and generate in one call; `wait_for_video` then returns once the video is done.
|
|
43
|
+
|
|
44
|
+
## Every operation is a method
|
|
45
|
+
|
|
46
|
+
Each operation of the [OpenAPI document](openapi/openapi.json) is a method named after its `operationId`:
|
|
47
|
+
`create_video`, `get_video`, `list_videos`, `quote_video`, `generate_video`, `clone_video`, `translate_video`,
|
|
48
|
+
`list_voices`, `list_templates`, `upload_file`, `create_upload_link`, `create_webhook`, `list_capabilities`,
|
|
49
|
+
`get_account`, and the rest.
|
|
50
|
+
|
|
51
|
+
- Path parameters are positional: `client.get_video("6f1c…")`.
|
|
52
|
+
- Query parameters and body fields are keyword arguments: `client.create_video(brief="…", auto_generate=True)`.
|
|
53
|
+
- Leave an argument out to not send it; pass `None` to send JSON `null`. Fields the SDK does not know yet go in
|
|
54
|
+
`extra_body={...}`.
|
|
55
|
+
- Answers are plain `dict`s, typed as `TypedDict`s in `zebracat.types` for your editor.
|
|
56
|
+
- Files (`upload_file`, `create_clone_voice`, `upload_to_link`) take bytes, an open binary file, or
|
|
57
|
+
`(filename, bytes, content_type)`.
|
|
58
|
+
- Operations the API marks deprecated still work and raise a `DeprecationWarning`.
|
|
59
|
+
|
|
60
|
+
## Uploading a file
|
|
61
|
+
|
|
62
|
+
`upload_file` takes a picture in one call. For a video to translate, clone or swap, or a voice sample, use a one-time
|
|
63
|
+
upload link: `upload_with_link` asks for the link (`create_upload_link`) and sends the file to it.
|
|
64
|
+
|
|
65
|
+
```python
|
|
66
|
+
stored = client.upload_with_link("talk.mp4", use="translate") # a local path, bytes or an open file
|
|
67
|
+
video = client.translate_video(video_url=stored["url"], target_language="es")
|
|
68
|
+
```
|
|
69
|
+
|
|
70
|
+
`use` is `picture`, `translate`, `clone`, `swap` or `voice_clone`; it sets the accepted types and the size cap.
|
|
71
|
+
|
|
72
|
+
## Authentication
|
|
73
|
+
|
|
74
|
+
```python
|
|
75
|
+
Zebracat(api_key="sk-...") # X-API-KEY, billed from the API balance
|
|
76
|
+
|
|
77
|
+
Zebracat( # OAuth: Authorization: Bearer zoa_..., billed from plan credits
|
|
78
|
+
access_token="zoa_...",
|
|
79
|
+
refresh_token="zor_...",
|
|
80
|
+
client_id="your-oauth-client-id",
|
|
81
|
+
client_secret=None, # only for confidential clients
|
|
82
|
+
on_token_refresh=lambda tokens: save(tokens),
|
|
83
|
+
)
|
|
84
|
+
```
|
|
85
|
+
|
|
86
|
+
With a refresh token, a `401` makes the client swap it for a new pair once (`POST /api/oauth/token`, the API rotates
|
|
87
|
+
both tokens) and repeat the call. `on_token_refresh` receives the new tokens so you can store them.
|
|
88
|
+
|
|
89
|
+
`ZEBRACAT_BASE_URL` (or `base_url=`) points the client at another server.
|
|
90
|
+
|
|
91
|
+
## Retries
|
|
92
|
+
|
|
93
|
+
A call is repeated up to `max_retries` times (default 3) when the API answers `429`, `500`, `502`, `503` or `504`,
|
|
94
|
+
or the connection drops:
|
|
95
|
+
|
|
96
|
+
- A `Retry-After` header is honoured. A wait longer than `max_retry_wait` (default 60 s, e.g. a daily cap) is raised
|
|
97
|
+
as `RateLimitError` with `retry_after` instead of slept through.
|
|
98
|
+
- Without `Retry-After` the wait is 0.5 s, 1 s, 2 s… (at most 8 s) with jitter.
|
|
99
|
+
- A `429` is always retried: the API refused before doing anything.
|
|
100
|
+
- A `5xx` or a dropped connection is retried only when resending cannot do the work twice: reads, `PUT`, `DELETE`,
|
|
101
|
+
and the paid POSTs whose server replays the first answer for a repeated `Idempotency-Key` (`create_video`,
|
|
102
|
+
`clone_video`, `translate_video` and the v1 creators). Other POSTs raise at once, so a charge is never repeated.
|
|
103
|
+
|
|
104
|
+
Per call: `request_options={"max_retries": 0, "timeout": 120, "headers": {...}}`.
|
|
105
|
+
|
|
106
|
+
## Idempotency keys
|
|
107
|
+
|
|
108
|
+
Every POST that spends money (it can answer `402`, or it buys credit) sends an `Idempotency-Key`. The SDK makes one
|
|
109
|
+
per call and reuses it across that call's retries. Pass your own to make a retry from your side safe too:
|
|
110
|
+
|
|
111
|
+
```python
|
|
112
|
+
client.create_video(brief="…", auto_generate=True, idempotency_key=f"order-{order_id}")
|
|
113
|
+
```
|
|
114
|
+
|
|
115
|
+
The same key with a different body answers `422 idempotency_key_reused`.
|
|
116
|
+
|
|
117
|
+
## Waiting for a video
|
|
118
|
+
|
|
119
|
+
`wait_for_video(task_id, timeout=1800, poll_interval=None, on_update=None)` reads the video again after the
|
|
120
|
+
`poll_after_s` seconds the API gives (falling back to 10–30 s per phase on an API that does not send it). It returns
|
|
121
|
+
when nothing more happens on its own:
|
|
122
|
+
|
|
123
|
+
| phase | what you do next |
|
|
124
|
+
|---|---|
|
|
125
|
+
| `completed` | `video["video_url"]` is the file; `download_video_file(task_id)` returns the URL as well |
|
|
126
|
+
| `plan_ready` | show `quote_video(task_id)`, then `generate_video(task_id)` and wait again |
|
|
127
|
+
| `needs_input` | answer `video["question"]` with `answer_video_question(task_id, ...)` and wait again |
|
|
128
|
+
|
|
129
|
+
It raises `VideoFailedError` (`.video`, `.phase`, `.error`) when the video fails or is cancelled, and
|
|
130
|
+
`WaitTimeoutError` (`.video`) when `timeout` passes first; the video keeps going on our side.
|
|
131
|
+
|
|
132
|
+
## Errors
|
|
133
|
+
|
|
134
|
+
Every error from the API is a subclass of `zebracat.APIError` and carries the public error envelope:
|
|
135
|
+
|
|
136
|
+
```python
|
|
137
|
+
from zebracat import InsufficientFundsError, RateLimitError, APIError
|
|
138
|
+
|
|
139
|
+
try:
|
|
140
|
+
client.generate_video(task_id)
|
|
141
|
+
except InsufficientFundsError as e:
|
|
142
|
+
print("Add credit:", e.topup_url)
|
|
143
|
+
except RateLimitError as e:
|
|
144
|
+
print("Try again in", e.retry_after, "s")
|
|
145
|
+
except APIError as e:
|
|
146
|
+
print(e.status_code, e.code, e.hint, e.request_id)
|
|
147
|
+
```
|
|
148
|
+
|
|
149
|
+
| status | class |
|
|
150
|
+
|---|---|
|
|
151
|
+
| 400, 422 | `BadRequestError` (`field`, `allowed_values`, `errors`) |
|
|
152
|
+
| 401 | `AuthenticationError` |
|
|
153
|
+
| 402 | `InsufficientFundsError` (`topup_url`) |
|
|
154
|
+
| 403 | `PermissionDeniedError` (`upgrade_url`) |
|
|
155
|
+
| 404, 410 | `NotFoundError` |
|
|
156
|
+
| 409 | `ConflictError` |
|
|
157
|
+
| 429 | `RateLimitError` (`retry_after`) |
|
|
158
|
+
| 5xx | `InternalServerError` |
|
|
159
|
+
| no answer | `APIConnectionError`, `APITimeoutError` |
|
|
160
|
+
|
|
161
|
+
`request_id` is the id to quote to support: the API's `request_id` (also the `X-Request-ID` header). The SDK sends
|
|
162
|
+
an `X-Request-ID` of its own on every call, so even a connection error has one.
|
|
163
|
+
|
|
164
|
+
## Development
|
|
165
|
+
|
|
166
|
+
```bash
|
|
167
|
+
uv venv && uv pip install -e '.[test]'
|
|
168
|
+
python scripts/generate.py # regenerate src/zebracat/_generated from openapi/openapi.json
|
|
169
|
+
python scripts/generate.py --check # CI: fails when the generated code is out of date
|
|
170
|
+
pytest # runs against a local stub server, no network
|
|
171
|
+
```
|
|
172
|
+
|
|
173
|
+
`openapi/openapi.json` comes from the backend: `python manage.py spectacular --format openapi-json --file
|
|
174
|
+
openapi.json` (or download `https://api.zebracat.ai/docs/openapi.json`). Copy it here, regenerate, commit both.
|
|
175
|
+
|
|
176
|
+
## Releasing
|
|
177
|
+
|
|
178
|
+
Bump `src/zebracat/_version.py`, commit, tag `vX.Y.Z` and push the tag. `.github/workflows/release.yml` tests,
|
|
179
|
+
builds and publishes to PyPI with trusted publishing (GitHub OIDC; no token is stored anywhere).
|