0.4.4

What's Changed

• Add logging.error for Retry logic by @jxnl in #259
• Update contributing.md by @tadash10 in #241
• Adds ruff linting and formatting by @ggaabe in #262
• Json mode nested models by @shanktt in #263
• Fix async streaming by @Anmol6 in #261
• Add multitask tests by @jxnl in #265
• Parse availability of dates by @jxnl in #267
• Integrate and Enhance Type Checking with Mypy by @gao-hongnan in #264
• Add new Mode for Anyscale's json schema schema type by @jxnl in #273
• Revert "Add new Mode for Anyscale's json schema schema type " by @jxnl in #274
• Support JSON_OBJECT mode from Anyscale by @jxnl in #275
• Updates to Tutorials by @jxnl in #248
• added finish reason exception IncompleteOutputException by @ionflow in #279
• doc: small change to readablity. by @Tedfulk in #281
• feat: add debugging for retries by @jxnl in #283
• blog: Third party models by @Anmol6 in #284

New Contributors

• @tadash10 made their first contribution in #241
• @ggaabe made their first contribution in #262
• @shanktt made their first contribution in #263
• @gao-hongnan made their first contribution in #264
• @ionflow made their first contribution in #279
• @Tedfulk made their first contribution in #281

Full Changelog: 0.4.2...0.4.3

0.4.2

What's Changed

• Opensource examples - Runpod w/Text-Generation-WebUI API Endpoint by @PhiBrandon in #247
• Fixed the system prompt for JSON mode, enabling the use of Pydantic nested models by @aastroza in #249
• Fix "seperate" typo by @toolittlecakes in #251
• added missing angle bracket by @toolittlecakes in #252
• Small doc fix for FastAPI section by @waseemhnyc in #253

New Contributors

• @toolittlecakes made their first contribution in #251
• @waseemhnyc made their first contribution in #253

Full Changelog: 0.4.1...0.4.2

0.4.1

What's Changed

• Fixed up missing dependencies and updated README.md by @ivanleomk in #232
• generator blog post by @Anmol6 in #230
• Update caching.md by @MeDott29 in #233
• Adding Open Router inference Example by @PhiBrandon in #226
• add: fastpi to concepts by @Muhtasham in #234
• Generator blog: small formatting by @Anmol6 in #237
• ref fastapi in generator.md by @Anmol6 in #240
• Opensource examples by @PhiBrandon in #238
• Fixed 1 spelling error in docs by @AlexTelon in #236
• chore: add grit pattern enforcement by @morgante in #243
• Fix #244 -- Refactor dump_message function to force content key by @Guiforge in #245
• Markdown JSON Mode by @Anmol6 in #246
• Documentation: Update planning-tasks.md by @taziksh in #250

New Contributors

• @PhiBrandon made their first contribution in #226
• @Muhtasham made their first contribution in #234
• @AlexTelon made their first contribution in #236
• @morgante made their first contribution in #243
• @Guiforge made their first contribution in #245
• @taziksh made their first contribution in #250

Full Changelog: 0.4.0...0.4.1

0.4.0

Many Improvements were made:

1. Multitask is swapped with Iterable[T]
2. More validations like content moderation
3. Supporting tool_call and JSON_MODE via the Mode patch.

What's Changed

• small fixes to tutorials by @fpingham in #168
• first version validation tutorial by @fpingham in #180
• Tutorials creative acts in documentation by @MeDott29 in #191
• Update index.md by @OxfordOutlander in #199
• Add coveralls by @jxnl in #203
• Ivan/tutorial cod by @ivanleomk in #193
• small fixes to tutorials by @fpingham in #205
• Minor corrections to jupyter notebooks by @aastroza in #206
• Openai mod validator by @fpingham in #207
• doc: add missing imports to some examples in README.md by @tavisrudd in #215
• Add multiple modalities: tools, functions, json_mode by @jxnl in #218
• Blog: Mastering Caching by @jxnl in #219
• Split coverage by @jxnl in #222
• Add client test fixtures by @jxnl in #227
• Support Streaming MultiTask with response_model by @Anmol6 in #221
• Add braintrust proxy by @ankrgyl in #225

New Contributors

• @OxfordOutlander made their first contribution in #199
• @aastroza made their first contribution in #206
• @tavisrudd made their first contribution in #215
• @Anmol6 made their first contribution in #221
• @ankrgyl made their first contribution in #225

Full Changelog: 0.3.5...0.4.0

0.3.5

What's Changed

• Spelling fix - chain-of-density.md by @mitch-36 in #176
• Fix llm_validator by @jxnl in #179
• Fix async usage by @Omegastick in #167
• Fix syntax error in code example by @smuotoe in #182
• Blog on learning some async options by @jxnl in #177
• improve documentation Readme by @jxnl in #186
• Correct typo by @daaniyaan in #170
• Creative acts in documentation by @MeDott29 in #188
• Improve Documentation by @jxnl in #189
• Add tutorials by @jxnl in #154
• Examples of using LLMs for citation verification by @jxnl in #192
• Update 2.tips.ipynb by @daaniyaan in #194
• RFC how much detail is too much? by @MeDott29 in #198

New Contributors

• @mitch-36 made their first contribution in #176
• @Omegastick made their first contribution in #167
• @smuotoe made their first contribution in #182

Full Changelog: 0.3.4...0.3.5

0.3.4

What's Changed

• Fix typo by @Xeophon in #173
• fixes retry_async message unpacking by @bradenkinard in #175

New Contributors

• @Xeophon made their first contribution in #173
• @bradenkinard made their first contribution in #175

Full Changelog: 0.3.3...0.3.4

0.3.3

What's Changed

• Added support for model suffix and added migrations for new OpenAI SDK by @ivanleomk in #169
• Chain of density by @ivanleomk in #135
• Chain of density edits by @ivanleomk in #171

Full Changelog: 0.3.2...0.3.3

0.3.2

What's Changed

• small typo fix by @bnkc in #160
• Fix apatch by @jxnl in #165

New Contributors

• @bnkc made their first contribution in #160

Full Changelog: 0.3.1...0.3.2

0.3.1

Full Changelog: 0.3.0...0.3.1

FIxed issues with classmethod

0.3.0

Upgrading to openai 1.1.0

Usage

from openai import OpenAI()
import instructor

# Enables `response_model`
client = instructor.patch(OpenAI())

class UserDetail(BaseModel):
    name: str
    age: int

user = client.chat.completions.create(
    model="gpt-3.5-turbo",
    response_model=UserDetail,
    messages=[
        {"role": "user", "content": "Extract Jason is 25 years old"},
    ]
)

assert isinstance(user, UserDetail)
assert user.name == "Jason"
assert user.age == 25

note "Using openai<1.0.0"

If you're using openai<1.0.0 then make sure you pip install instructor<0.3.0
where you can patch a global client like so:

import openai
import instructor

instructor.patch()

user = openai.ChatCompletion.create(
    ...,
    response_model=UserDetail,
)

What's Changed

• Migration to OpenAI 1.1.0 by @grit-app in #152

New Contributors

• @grit-app made their first contribution in #152

Full Changelog: 0.2.11...0.3.0