Metadata-Version: 2.4
Name: sriti-core
Version: 0.1.1
Summary: Adaptive LLM cascade engine with semantic caching and reliability tracking
Author: Sudev Mukherjee, Purusoth Mahendran, Clint Solomon
License:                                  Apache License
                                   Version 2.0, January 2004
                                http://www.apache.org/licenses/
        
           TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
        
           1. Definitions.
        
              "License" shall mean the terms and conditions for use, reproduction,
              and distribution as defined by Sections 1 through 9 of this document.
        
              "Licensor" shall mean the copyright owner or entity authorized by
              the copyright owner that is granting the License.
        
              "Legal Entity" shall mean the union of the acting entity and all
              other entities that control, are controlled by, or are under common
              control with that entity. For the purposes of this definition,
              "control" means (i) the power, direct or indirect, to cause the
              direction or management of such entity, whether by contract or
              otherwise, or (ii) ownership of fifty percent (50%) or more of the
              outstanding shares, or (iii) beneficial ownership of such entity.
        
              "You" (or "Your") shall mean an individual or Legal Entity
              exercising permissions granted by this License.
        
              "Source" form shall mean the preferred form for making modifications,
              including but not limited to software source code, documentation
              source, and configuration files.
        
              "Object" form shall mean any form resulting from mechanical
              transformation or translation of a Source form, including but
              not limited to compiled object code, generated documentation,
              and conversions to other media types.
        
              "Work" shall mean the work of authorship made available under
              the License, as indicated by a copyright notice that is included in
              or attached to the work (an example is provided in the Appendix below).
        
              "Derivative Works" shall mean any work, whether in Source or Object
              form, that is based on (or derived from) the Work and for which the
              editorial revisions, annotations, elaborations, or other modifications
              represent, as a whole, an original work of authorship. For the purposes
              of this License, Derivative Works shall not include works that remain
              separable from, or merely link (or bind by name) to the interfaces of,
              the Work and Derivative Works thereof.
        
              "Contribution" shall mean, as submitted to the Licensor for inclusion
              in the Work by the copyright owner or by an individual or Legal Entity
              authorized to submit on behalf of the copyright owner. For the purposes
              of this definition, "submitted" means any form of electronic, verbal,
              or written communication sent to the Licensor or its representatives,
              including but not limited to communication on electronic mailing lists,
              source code control systems, and issue tracking systems that are managed
              by, or on behalf of, the Licensor for the purpose of discussing and
              improving the Work, but excluding communication that is conspicuously
              marked or designated in writing by the copyright owner as "Not a
              Contribution."
        
              "Contributor" shall mean Licensor and any Legal Entity on behalf of
              whom a Contribution has been received by the Licensor and included
              within the Work.
        
           2. Grant of Copyright License. Subject to the terms and conditions of
              this License, each Contributor hereby grants to You a perpetual,
              worldwide, non-exclusive, no-charge, royalty-free, irrevocable
              copyright license to reproduce, prepare Derivative Works of,
              publicly display, publicly perform, sublicense, and distribute the
              Work and such Derivative Works in Source or Object form.
        
           3. Grant of Patent License. Subject to the terms and conditions of
              this License, each Contributor hereby grants to You a perpetual,
              worldwide, non-exclusive, no-charge, royalty-free, irrevocable
              (except as stated in this section) patent license to make, have made,
              use, offer to sell, sell, import, and otherwise transfer the Work,
              where such license applies only to those patent claims licensable
              by such Contributor that are necessarily infringed by their
              Contribution(s) alone or by the combined work of their Contribution(s)
              with the Work to which such Contribution(s) was submitted. If You
              institute patent litigation against any entity (including a cross-claim
              or counterclaim in a lawsuit) alleging that the Work or any
              Contribution embodied within the Work constitutes direct or contributory
              patent infringement, then any patent licenses granted to You under
              this License for that Work shall terminate as of the date such
              litigation is filed.
        
           4. Redistribution. You may reproduce and distribute copies of the
              Work or Derivative Works thereof in any medium, with or without
              modifications, and in Source or Object form, provided that You
              meet the following conditions:
        
              (a) You must give any other recipients of the Work or Derivative
                  Works a copy of this License; and
        
              (b) You must cause any modified files to carry prominent notices
                  stating that You changed the files; and
        
              (c) You must retain, in the Source form of any Derivative Works
                  that You distribute, all copyright, patent, trademark, and
                  attribution notices from the Source form of the Work,
                  excluding those notices that do not pertain to any part of
                  the Derivative Works; and
        
              (d) If the Work includes a "NOTICE" text file as part of its
                  distribution, You must include a readable copy of the
                  attribution notices contained within such NOTICE file, in
                  at least one of the following places: within a NOTICE text
                  file distributed as part of the Derivative Works; within
                  the Source form or documentation, if provided along with the
                  Derivative Works; or, within a display generated by the
                  Derivative Works, if and wherever such third-party notices
                  normally appear. The contents of the NOTICE file are for
                  informational purposes only and do not modify the License.
                  You may add Your own attribution notices within Derivative
                  Works that You distribute, alongside or in addition to the
                  NOTICE text from the Work, provided that such additional
                  attribution notices cannot be construed as modifying the License.
        
              You may add Your own license statement for Your modifications and
              may provide additional grant of rights to use, copy, modify, merge,
              publish, distribute, sublicense, and/or sell copies of the
              Contribution, either on its own or together with this Work.
        
           5. Submission of Contributions. Unless You explicitly state otherwise,
              any Contribution intentionally submitted for inclusion in the Work
              by You to the Licensor shall be under the terms and conditions of
              this License, without any additional terms or conditions.
              Notwithstanding the above, nothing herein shall supersede or modify
              the terms of any separate license agreement you may have executed
              with Licensor regarding such Contributions.
        
           6. Trademarks. This License does not grant permission to use the trade
              names, trademarks, service marks, or product names of the Licensor,
              except as required for reasonable and customary use in describing the
              origin of the Work and reproducing the content of the NOTICE file.
        
           7. Disclaimer of Warranty. Unless required by applicable law or
              agreed to in writing, Licensor provides the Work (and each
              Contributor provides its Contributions) on an "AS IS" BASIS,
              WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
              implied, including, without limitation, any warranties or conditions
              of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
              PARTICULAR PURPOSE. You are solely responsible for determining the
              appropriateness of using or reproducing the Work and assume any
              risks associated with Your exercise of permissions under this License.
        
           8. Limitation of Liability. In no event and under no legal theory,
              whether in tort (including negligence), contract, or otherwise,
              unless required by applicable law (such as deliberate and grossly
              negligent acts) or agreed to in writing, shall any Contributor be
              liable to You for damages, including any direct, indirect, special,
              incidental, or exemplary damages of any character arising as a
              result of this License or out of the use or inability to use the
              Work (including but not limited to damages for loss of goodwill,
              work stoppage, computer failure or malfunction, or all other
              commercial damages or losses), even if such Contributor has been
              advised of the possibility of such damages.
        
           9. Accepting Warranty or Additional Liability. While redistributing
              the Work or Derivative Works thereof, You may choose to offer,
              and charge a fee for, acceptance of support, warranty, indemnity,
              or other liability obligations and/or rights consistent with this
              License. However, in accepting such obligations, You may offer only
              conditions consistent with this License.
        
           END OF TERMS AND CONDITIONS
        
           Copyright 2024 Sudev Mukherjee, Purusoth Mahendran, Clint Solomon
        
           Licensed under the Apache License, Version 2.0 (the "License");
           you may not use this file except in compliance with the License.
           You may obtain a copy of the License at
        
               http://www.apache.org/licenses/LICENSE-2.0
        
           Unless required by applicable law or agreed to in writing, software
           distributed under the License is distributed on an "AS IS" BASIS,
           WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
           implied. See the License for the specific language governing
           permissions and limitations under the License.
        
Project-URL: Homepage, https://github.com/sriti-ai/sriti-core
Project-URL: Issues, https://github.com/sriti-ai/sriti-core/issues
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.11
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: fastapi>=0.115.0
Requires-Dist: uvicorn[standard]>=0.30.0
Requires-Dist: httpx>=0.27.2
Requires-Dist: pydantic<3.0,>=2.0
Requires-Dist: pydantic-settings>=2.5.2
Requires-Dist: python-dotenv>=1.0.1
Requires-Dist: litellm<1.88,>=1.83.7
Requires-Dist: redis<6,>=5.0.8
Requires-Dist: pyyaml>=6.0.2
Requires-Dist: numpy>=1.26.0
Requires-Dist: fastembed>=0.3.0
Requires-Dist: llmlingua>=0.2.0
Requires-Dist: torch>=2.8.0
Requires-Dist: transformers>=4.53.0
Requires-Dist: cryptography>=42.0.0
Provides-Extra: dev
Requires-Dist: pytest>=8.0; extra == "dev"
Requires-Dist: pytest-asyncio>=0.24; extra == "dev"
Dynamic: license-file

<p align="center">
  <img src="./assets/logo.png" alt="Sriti Core Logo" width="220" />
</p>

<h1 align="center">Sriti Core</h1>

<p align="center">
  <strong>Intelligent, Reliability-Aware Model Routing & Cascade Engine for LLMs</strong>
</p>

<p align="center">
  <a href="#key-features">Key Features</a> •
  <a href="#architecture">Architecture</a> •
  <a href="#quickstart">Quickstart</a> •
  <a href="#how-it-works">How It Works</a> •
  <a href="#configuration">Configuration</a> •
  <a href="#api-reference">API Reference</a> •
  <a href="#testing">Testing</a> •
  <a href="#troubleshooting">Troubleshooting</a> •
  <a href="#license">License</a>
</p>

---

## Overview

**Sriti Core** is an open-source, ultra-efficient intelligent routing proxy and cascading engine for Large Language Models.

Instead of routing every single request to expensive frontier models (like Claude 3.5 Sonnet or GPT-4o), Sriti dynamically categorizes the request, checks an encrypted semantic cache, and cascades through a hierarchy of models:
1. **Tier 3 (Local / Fast):** Free, local models running on edge devices or on-prem hardware via Ollama (e.g., Qwen 2.5, Llama 3.2, MiniCPM).
2. **Tier 2 (Balanced Cloud):** High-throughput, cost-effective inference providers (e.g., Groq, Fireworks, DeepInfra).
3. **Tier 1 (Frontier):** Capable reasoning models invoked only when task complexity warrants it, or when lower tiers fail verification or latency SLOs.

Sriti continuously learns model reliability, evaluates output quality gates, and guarantees strict latency SLOs while dramatically slashing cloud API bills.

---

## Key Features

- 🎯 **In-Process Task Classifier:** Zero-shot task classification using fast ONNX runtime (`fastembed` BAAI/bge-small-en-v1.5) mapping prompts across categories (e.g., Code, Extraction, Summarization, QA, Math, Reasoning).
- ⚡ **Semantically Aware Caching:** Vector similarity caching backed by Redis/Valkey with task-dependent cosine thresholds. Uncacheable tasks (e.g., non-deterministic reasoning, math) bypass the cache automatically.
- 🔒 **Encrypted Cache at Rest:** Optional AES-GCM (256-bit) authenticated encryption for cached prompt and completion payloads.
- 🗜️ **In-Process Prompt Compression:** LLMLingua-2 integration to compress lengthy contexts before sending them over the wire, with an adaptive quality gate to ensure semantic integrity.
- 📊 **Dynamic Cascade with Quality Gates:** Evaluates lower-tier completions using structural validation (e.g. JSON validation) or semantic quality scoring. If a tier's response fails, it automatically escalates to the next tier.
- 📈 **Online Reliability Profile Learner:** Real-time Exponential Moving Average (EMA) tracking of per-model latency, success rate, and error states without needing external analytical databases.
- 🧠 **Case-Based Memory (KNN Routing):** Remembers past routing decisions and dynamically adjusts candidate selection bias based on historical outcome rewards.
- 🛡️ **Lightweight & Self-Contained:** Uses a native NumPy cosine similarity vector store requiring no proprietary Redis modules (no RediSearch or RedisJSON needed). Works seamlessly on vanilla Redis or Valkey.
- 🔌 **Unified LLM Execution:** Powered by LiteLLM to seamlessly talk to 100+ LLM providers and local inference runtimes.

---

## Architecture

```text
               User / Application Request
                           │
                           ▼
          ┌──────────────────────────────────┐
          │     1. Fast ONNX Classifier      │ ──> Task Category & Embeddings
          └──────────────────────────────────┘
                           │
                           ▼
          ┌──────────────────────────────────┐
          │    2. Semantic Cache (Valkey)    │ ──[Hit]──> Return Cached Response
          └──────────────────────────────────┘
                           │ [Miss]
                           ▼
          ┌──────────────────────────────────┐
          │  3. Model Cascading Engine       │
          │                                  │
          │  ┌────────────────────────────┐  │
          │  │ Tier 3: Local Edge Models  │  │ ──[Success]──> Accept & Learn
          │  └────────────────────────────┘  │
          │                 │ [Fail/Timeout] │
          │  ┌────────────────────────────┐  │
          │  │ Tier 2: Cost-Effective LLM │  │ ──[Success]──> Accept & Learn
          │  └────────────────────────────┘  │
          │                 │ [Fail/Timeout] │
          │  ┌────────────────────────────┐  │
          │  │ Tier 1: Frontier Models    │  │ ──[Guaranteed Execution]
          │  └────────────────────────────┘  │
          └──────────────────────────────────┘
                           │
                           ▼
          ┌──────────────────────────────────┐
          │  4. Update Reliability & Cache   │
          └──────────────────────────────────┘
```

---

## Quickstart

### 1. Installation

Python 3.11+ is required.

**Via pip (recommended):**
```bash
pip install sriti-core
```

**From source:**
```bash
git clone https://github.com/sriti-ai/sriti-core.git
cd sriti-core
pip install -e ".[dev]"
```

### 2. Configure Environment

Create a `.env` file with your credentials and configuration:

```env
SRITI_REDIS_URL=redis://localhost:6379/0
# Optional: Provider API keys (only configure what you use)
ANTHROPIC_API_KEY=sk-ant-...
OPENAI_API_KEY=sk-...
GROQ_API_KEY=gsk_...
# Optional: AES-256 GCM key for cache encryption (base64-encoded 32 bytes)
# CACHE_ENCRYPTION_KEY=...
```

### 3. Start the Sriti Server

```bash
uvicorn sriti.app:app --host 0.0.0.0 --port 8100
```

### 4. Send a Request via HTTP

```bash
curl -X POST http://localhost:8100/complete \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Summarise the following paragraph in one sentence: ...",
    "minimum_tier": "tier_3",
    "cacheable": true,
    "requires_json": false
  }'
```

### 5. Python Client SDK

You can also use the bundled async Python client:

```python
import asyncio
from sriti.client import SritiModelClient

async def main():
    client = SritiModelClient(base_url="http://localhost:8100")
    response = await client.complete(
        prompt="Explain how semantic caching reduces LLM inference costs.",
        minimum_tier="tier_3",
        cacheable=True
    )
    print(f"Response: {response.text}")
    print(f"Tier Used: {response.tier_used}")
    print(f"Latency: {response.latency_ms}ms")
    print(f"Cost: ${response.cost_usd}")

asyncio.run(main())
```

---

## How It Works

### 1. Task Classification
Prompts are embedded using `fastembed` and compared against predefined semantic task anchors (or an optional lightweight classification head). The classifier runs completely locally in under 10ms.

### 2. Multi-Tier Semantic Cache
- Tasks have configurable similarity thresholds in `policy.yaml`. For example, conversational questions allow a similarity threshold of `0.75`, while structured data extraction requires `0.92`.
- Vectors and encrypted payloads are stored in Valkey/Redis.

### 3. Reliability-Driven Cascading
- For each tier, active candidates in `models.yaml` are evaluated based on their Pareto score: a composite metric factoring in latency, cost, and historical success probability.
- If a lower tier experiences an API error, context window exhaustion, or fails a verification check (e.g. invalid JSON syntax), Sriti escalates seamlessly to the next tier.

---

## Configuration

Sriti's configuration is managed through two clean YAML files in `sriti/config/`:

- **`models.yaml`**: Defines model providers, context windows, cost per token, and tier mappings (`default_tier: 1 | 2 | 3`).
- **`policy.yaml`**: Defines cache TTLs, per-task similarity thresholds, latency SLOs, and quality gate cutoffs.

---

## Testing

Run unit and cascade integration tests using pytest:

```bash
python -m pytest
```

All 39 unit tests run without requiring live LLM API keys or a running Redis instance (mocked in-memory).

---

## Troubleshooting

### `uvicorn: command not found`
The `uvicorn` binary is not on your PATH. Run it via Python instead:
```bash
python -m uvicorn sriti.app:app --host 0.0.0.0 --port 8100
```

### `ModuleNotFoundError: No module named 'fastapi'` (or other missing modules)
You installed `sriti-core` but the server dependencies are not in your active environment. Install them:
```bash
pip install sriti-core[dev]
# or from source:
pip install -e ".[dev]"
```

### `Could not find a version that satisfies the requirement sriti-core (from versions: none)`
Your Python version is below 3.11. Verify:
```bash
python --version
```
sriti-core requires Python 3.11+. If your default Python is older (e.g. Anaconda 3.8), create a compatible environment:
```bash
conda create -n sriti-env python=3.11 -y
conda activate sriti-env
pip install sriti-core
```

### `pip install -e ".[dev]"` fails with `setup.py not found`
Your pip version is too old to support `pyproject.toml`-based editable installs. Upgrade pip first:
```bash
pip install --upgrade pip
pip install -e ".[dev]"
```

### `pytest` collects 0 items or fails with `ModuleNotFoundError`
Your shell is picking up the system/Anaconda pytest instead of the one in your active environment. Use:
```bash
python -m pytest
```

### `Could not find a version that satisfies the requirement torch>=2.8.0`
PyTorch 2.8+ requires macOS 13 (Ventura) or later and a recent pip. If you are on macOS 12 or older, you cannot install sriti-core from PyPI directly. Options:

- **Upgrade macOS** to 13+ (recommended for full compatibility).
- **Run on Linux** — all dependencies resolve cleanly on modern Ubuntu/Debian.
- **Use Docker** — run the server in a container with a compatible base image:
  ```bash
  docker run --rm -it python:3.11-slim bash
  pip install sriti-core
  ```

### macOS: wrong Python is used even after `conda activate`
If `which python` still points to Anaconda's base Python, your shell may not have conda initialized. Run:
```bash
source /opt/anaconda3/etc/profile.d/conda.sh
conda activate sriti-env
which python   # should now show the sriti-env path
```

---

## License

This project is licensed under the Apache License 2.0.
