jharness-models

OpenAI Chat Completions, OpenAI Responses, and Anthropic Messages adapters; DeepSeek profiles; and provider-neutral model composition for the JHarness kernel.

uv add jharness-models

from jharness.models.openai import OpenAIResponsesModel

Adapter	Runtime tools	Provider-hosted tools	Ordered output
OpenAI Chat Completions	Function calls	None	Content and calls are normalized into `ModelResponse.output`
Anthropic Messages	Client `tool_use`	Profile-installed server-tool codecs	Native block order is retained
OpenAI Responses	Function and freeform calls	Profile-installed image generation and web search codecs	Native Responses item order is retained

DeepSeek's native Responses endpoint uses OpenAIResponsesModel with deepseek_openai_responses_profile. That profile is text-only, accepts only deepseek-v4-flash, exposes provider-hosted web search plus the exact freeform apply_patch runtime tool, and forces stateless requests with complete history. The DeepSeek Anthropic profile independently exposes its verified server-side web search.

Model modalities describe what the model itself understands or produces. Tool ownership is separate: RuntimeToolCall is executed by the JHarness runtime, while a ProviderToolCall records work already executed by the supplier. Both remain interleaved with ContentPart values in ordered output.

Each protocol profile contains the exact immutable ModelCapabilities returned by its model client. The default Responses profile is conservative and text-only; model-specific modalities, structured output, and hosted tools require explicit host opt-in. Tool selection is declared as a set of supported types rather than a coarse boolean. Supplier factories only compose protocol capabilities and wire policies; the shared codecs contain no supplier-name branches.

OpenAI Responses sends store=false by default and requests encrypted reasoning history. Hosted image generation additionally requires a host-owned ResponsesArtifactStore; generated base64 is persisted externally and durable history contains only integrity-bearing ArtifactRef values. Stores must be durable, idempotent, safe for provider-controlled call ids, available during run recovery, and responsible for retention and garbage collection of uncommitted saves.

Retry and fallback compose directly around model values:

from jharness.models.decorators import FallbackModel, RetryingModel

model = FallbackModel(
    RetryingModel(primary_model, max_attempts=3),
    RetryingModel(backup_model, max_attempts=2),
)

Installing this distribution installs the exact matching jharness-kernel version. Provider configuration and composition details are in the model adapter guide.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

jharness_models-0.6.0.tar.gz (63.5 kB view details)

Uploaded Aug 12, 2026 Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

jharness_models-0.6.0-py3-none-any.whl (83.9 kB view details)

Uploaded Aug 12, 2026 Python 3

File details

Details for the file jharness_models-0.6.0.tar.gz.

File metadata

Download URL: jharness_models-0.6.0.tar.gz
Upload date: Aug 12, 2026
Size: 63.5 kB
Tags: Source
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for jharness_models-0.6.0.tar.gz
Algorithm	Hash digest
SHA256	`1f088302158c79e2e5a641e1f49cd9b124c14f211e33eaff2ca6893f733daf57`
MD5	`4e65611c0e873a9289e39a057c6e951e`
BLAKE2b-256	`8428d793507e32a1192875e721c39df59a3d86cc8d71034d2621890f6c8f750d`

See more details on using hashes here.

Provenance

The following attestation bundles were made for jharness_models-0.6.0.tar.gz:

Publisher: release.yml on Ezio2000/jharness

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Statement:
- Statement type: https://in-toto.io/Statement/v1
- Predicate type: https://docs.pypi.org/attestations/publish/v1
- Subject name: jharness_models-0.6.0.tar.gz
- Subject digest: 1f088302158c79e2e5a641e1f49cd9b124c14f211e33eaff2ca6893f733daf57
- Sigstore transparency entry: 2432913033
- Sigstore integration time: Aug 12, 2026
Source repository:
- Permalink: Ezio2000/jharness@438bf4434a98e013c13fabf5af0e620119dc5c65
- Branch / Tag: refs/heads/main
- Owner: https://github.com/Ezio2000
- Access: public
Publication detail:
- Token Issuer: https://token.actions.githubusercontent.com
- Runner Environment: github-hosted
- Publication workflow: release.yml@438bf4434a98e013c13fabf5af0e620119dc5c65
- Trigger Event: workflow_dispatch