You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+68-89Lines changed: 68 additions & 89 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -6,14 +6,12 @@
6
6
The Parallel Python library provides convenient access to the Parallel REST API from any Python 3.9+
7
7
application. The library includes type definitions for all request params and response fields,
8
8
and offers both synchronous and asynchronous clients powered by [httpx](https://github.com/encode/httpx).
9
-
It is strongly encouraged to use the asynchronous client for best performance.
10
9
11
10
It is generated with [Stainless](https://www.stainless.com/).
12
11
13
12
## Documentation
14
13
15
-
The REST API documentation can be found in our [docs](https://docs.parallel.ai).
16
-
The full API of this Python library can be found in [api.md](api.md).
14
+
The REST API documentation can be found on [docs.parallel.ai](https://docs.parallel.ai). The full API of this library can be found in [api.md](api.md).
To get the best performance out of Parallel's API, we recommend
80
-
using the asynchronous client, especially for executing multiple Task Runs concurrently.
81
-
Functionality between the synchronous and asynchronous clients is identical, including
82
-
the convenience methods.
72
+
Functionality between the synchronous and asynchronous clients is otherwise identical.
83
73
84
-
##Frequently Asked Questions
74
+
### With aiohttp
85
75
86
-
**Does the Task API accept prompts or objectives?**
76
+
By default, the async client uses `httpx` for HTTP requests. However, for improved concurrency performance you may also use `aiohttp` as the HTTP backend.
87
77
88
-
No, there are no `objective` or `prompt` parameters that can be specified for calls to
89
-
the Task API. Instead, provide any directives or instructions via the schemas. For
90
-
more information, check [our docs](https://docs.parallel.ai/task-api/core-concepts/specify-a-task).
78
+
You can enable this by installing `aiohttp`:
91
79
92
-
**Can I access beta parameters or endpoints via the SDK?**
80
+
```sh
81
+
# install from PyPI
82
+
pip install parallel-web[aiohttp]
83
+
```
93
84
94
-
Yes, the SDK supports both beta endpoints and beta header parameters for the Task API.
95
-
All beta parameters are accessible via the `client.beta` namespace in the SDK.
85
+
Then you can enable it by instantiating the client with `http_client=DefaultAioHttpClient()`:
96
86
97
-
**Can I specify a timeout for API calls?**
87
+
```python
88
+
import os
89
+
import asyncio
90
+
from parallel import DefaultAioHttpClient
91
+
from parallel import AsyncParallel
98
92
99
-
Yes, all methods support a timeout. For more information, see [Timeouts](#timeouts).
100
93
94
+
asyncdefmain() -> None:
95
+
asyncwith AsyncParallel(
96
+
api_key=os.environ.get("PARALLEL_API_KEY"), # This is the default and can be omitted
97
+
http_client=DefaultAioHttpClient(),
98
+
) as client:
99
+
task_run =await client.task_run.create(
100
+
input="What was the GDP of France in 2023?",
101
+
processor="base",
102
+
)
103
+
print(task_run.interaction_id)
104
+
105
+
106
+
asyncio.run(main())
107
+
```
108
+
109
+
## Using types
101
110
102
-
**Can I specify retries via the SDK?**
111
+
Nested request parameters are [TypedDicts](https://docs.python.org/3/library/typing.html#typing.TypedDict). Responses are [Pydantic models](https://docs.pydantic.dev) which also provide helper methods for things like:
103
112
104
-
Yes, errors can be retried via the SDK — the default retry count is 2. The maximum number
105
-
of retries can be configured at the client level. For information on which errors
106
-
are automatically retried and how to configure retry settings, see [Retries](#retries).
113
+
- Serializing back into JSON, `model.to_json()`
114
+
- Converting to a dictionary, `model.to_dict()`
107
115
108
-
## Low‑level API access
116
+
Typed requests and responses provide autocomplete and documentation within your editor. If you would like to see type errors in VS Code to help catch bugs earlier, set `python.analysis.typeCheckingMode` to `basic`.
109
117
110
-
The library also provides low‑level access to the Parallel API.
118
+
## Nested params
119
+
120
+
Nested parameters are dictionaries, typed using `TypedDict`, for example:
111
121
112
122
```python
113
123
from parallel import Parallel
114
-
from parallel.types import TaskSpecParam
115
124
116
125
client = Parallel()
117
126
118
127
task_run = client.task_run.create(
119
-
input={"country": "France", "year": 2023},
120
-
processor="core",
121
-
task_spec={
122
-
"output_schema": {
123
-
"json_schema": {
124
-
"additionalProperties": False,
125
-
"properties": {
126
-
"gdp": {
127
-
"description": "GDP in USD for the year",
128
-
"type": "string",
129
-
}
130
-
},
131
-
"required": ["gdp"],
132
-
"type": "object",
133
-
},
134
-
"type": "json",
135
-
},
136
-
"input_schema": {
137
-
"json_schema": {
138
-
"additionalProperties": False,
139
-
"properties": {
140
-
"country": {
141
-
"description": "Name of the country to research",
142
-
"type": "string",
143
-
},
144
-
"year": {
145
-
"description": "Year for which to retrieve information",
When the library is unable to connect to the API (for example, due to network connection problems or a timeout), a subclass of `parallel.APIConnectionError` is raised.
@@ -179,7 +148,10 @@ from parallel import Parallel
0 commit comments