nova-prompt-optimizer 1.0.48__tar.gz

nova_prompt_optimizer-1.0.48/LICENSE

@@ -0,0 +1,175 @@
+
+                                 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, whether in Source or
+      Object form, 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 any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to 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 otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated 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 combination 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 a Contribution incorporated 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, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, 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 as an addendum 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 copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   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 redistributing 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 consequential 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 any and 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 act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.

nova_prompt_optimizer-1.0.48/NOTICE

@@ -0,0 +1 @@
+Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.

nova_prompt_optimizer-1.0.48/PKG-INFO

@@ -0,0 +1,356 @@
+Metadata-Version: 2.4
+Name: nova-prompt-optimizer
+Version: 1.0.48
+Summary: A Python SDK for optimizing prompts for Nova.
+Author-email: Amazon <nova-prompt-optimizer@amazon.com>
+License-Expression: Apache-2.0
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+License-File: NOTICE
+Requires-Dist: jinja2
+Requires-Dist: boto3
+Requires-Dist: botocore
+Requires-Dist: boto3-stubs
+Requires-Dist: dspy
+Requires-Dist: numpy==2.3.1
+Requires-Dist: virtualenv==20.31.2
+Requires-Dist: urllib3==2.5.0
+Requires-Dist: setuptools==80.9.0
+Requires-Dist: h11==0.16.0
+Provides-Extra: dev
+Requires-Dist: build; extra == "dev"
+Requires-Dist: setuptools; extra == "dev"
+Requires-Dist: twine; extra == "dev"
+Requires-Dist: wheel; extra == "dev"
+Dynamic: license-file
+
+# Nova Prompt Optimizer
+
+A Python SDK for optimizing prompts for Nova.
+
+## 📚 Table of contents
+* [Installation](#installation)
+* [Pre-Requisites](#pre-requisites)
+* [Quick Start: Facility Support Analyzer Dataset](#-quick-start)
+* [Core Concepts](#core-concepts)
+  * [Input Adapters](#input-adapters)
+    * [1. Prompt Adapter](#1-prompt-adapter)
+    * [2. Inference Adapter](#2-inference-adapter)
+    * [3. Dataset Adapter](#3-dataset-adapter)
+    * [4. Metric Adapter](#4-metric-adapter)
+    * [5. Optimization Adapter](#5-optimization-adapter)
+  * [Optimizers](#optimizers)
+    * [NovaPromptOptimizer](#novapromptoptimizer)
+  * [Evaluator](#evaluator)
+* [Optimization Recommendations](#optimization-recommendations)
+* [Preview Status](#-preview-status)
+* [Acknowledgements](#acknowledgements)
+
+## Installation
+
+Install the library using
+```sh
+pip install amzn-nova-prompt-optimizer
+```
+
+
+## Pre-Requisites
+
+#### Setup your AWS Access Keys:
+
+To execute the SDK, you will need AWS credentials configured. Take a look at the [AWS CLI configuration documentation](https://docs.aws.amazon.com/cli/latest/userguide/cli-chap-configure.html#config-settings-and-precedence) for details on the various ways to configure credentials. 
+An easy way to try out the SDK is to populate the following environment variables with your AWS API credentials. 
+Take a look at this guide for [Authenticating with short-term credentials for the AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/cli-authentication-short-term.html)
+
+```commandline
+export AWS_ACCESS_KEY_ID="..."
+export AWS_SECRET_ACCESS_KEY="..."
+```
+
+#### To enable Nova model access:
+1. Go to the Amazon Bedrock Model Access page
+2. Click "Manage model access"
+3. Choose Amazon as provider and Nova models
+4. Click "Request access"
+5. Wait for approval (instant in most cases)
+
+
+## 🏁 Quick Start
+### Facility Support Analyzer Dataset
+The Facility Support Analyzer dataset consists of emails that are to be classified based on category, urgency and sentiment.
+
+Please see the [samples](samples/facility-support-analyzer/) folder for example notebooks of how to optimize a prompt in the scenario where a [user prompt template is to be optimized](samples/facility-support-analyzer/user_prompt_only) and the scenario where a [user and system prompt is to be optimized together](samples/facility-support-analyzer/system_and_user_prompt)
+
+
+## Core Concepts
+
+### Input Adapters
+We break each component of Prompt Optimization into Adapters providing a modular approach to Prompt Optimization.
+
+![adapters](docs/adapters.png)
+
+### 1. Prompt Adapter
+
+**Responsibility:** Ability to load prompts from different formats and store them in the standardized format (JSON)
+
+
+**Sample Prompt Adapter Initialization**
+
+```python
+from amzn_nova_prompt_optimizer.core.input_adapters.prompt_adapter import TextPromptAdapter
+
+prompt_adapter = TextPromptAdapter()
+
+prompt_adapter.set_system_prompt(file_path="prompt/sys_prompt.txt", variables={"foo"})
+
+prompt_adapter.set_user_prompt(content="You are a .....", variables={"bar"})
+
+prompt_adapter.adapt()
+```
+
+**Supported Prompt Adapters:** `TextPromptAdapter`
+
+Learn More about the Prompt Adapter [here](docs/PromptAdapter.md)
+
+### 2. Inference Adapter
+**Responsibility:** Ability to call an inference backend for the models e.g. Bedrock, etc.
+
+**Sample use of Inference Adapter**
+
+```python
+from amzn_nova_prompt_optimizer.core.inference.adapter import BedrockInferenceAdapter
+
+inference_adapter = BedrockInferenceAdapter(region_name="us-east-1")
+```
+
+You can pass `rate_limit` into constructor of InferenceAdapter to limit the max TPS of bedrock call to avoid throttle. Default to 2 if not set.
+
+```python
+from amzn_nova_prompt_optimizer.core.inference.adapter import BedrockInferenceAdapter
+
+inference_adapter = BedrockInferenceAdapter(region_name="us-east-1", rate_limit=10) # Max 10 TPS
+```
+
+**Supported Inference Adapters:** `BedrockInferenceAdapter`
+
+**Core Functions**
+
+Call the model using the parameters
+```python
+# Call the model with the passed parametrs
+inference_output = inference_adapter.call_model(model_id: str, system_prompt: str, messages: List[Dict[str, str]], inf_config: Dict[str, Any])
+```
+
+The Inference Adapter accepts the `system_prompt` as a string.
+
+The input to the model as a list of User/Assistant turns (messages). e.g. `[{"user": "foo"}, {"assistant": "bar"}, {"user": "What comes next?"}]`
+
+### 3. Dataset Adapter
+
+**Responsibility:** Ability to read/write datasets from different formats. Uses an intermediary standardized format when communicating with other adapters.
+It can also read list of JSON object. It can also create Train/Test splits (with stratify capability if set).
+
+**Requirements:** Currently, you can only provide a singleton set as output column.
+
+**Sample Dataset Adapter Initialization**
+```python
+# Example Usage
+from amzn_nova_prompt_optimizer.core.input_adapters.dataset_adapter import JSONDatasetAdapter
+
+input_columns = {"input"}
+output_columns = {"answer"}
+
+dataset_adapter = JSONDatasetAdapter(input_columns, output_columns)
+
+# Adapt
+dataset_adapter.adapt(data_source="sample_data.jsonl")
+
+# Split
+train, test = dataset_adapter.split(0.5)
+```
+
+**Supported Dataset Adapters:** `JSONDatasetAdapter`, `CSVDatasetAdapter`
+
+Learn More about the Dataset Adapter [here](docs/DatasetAdapter.md)
+
+### 4. Metric Adapter
+**Responsibility:** Ability to load custom metrics and apply them on inference output and ground truth
+
+**Metric Adapter Class**
+```python
+class MetricAdapter():
+    def apply(self, y_pred: Any, y_true: Any) -> float:
+        pass
+
+    def batch_apply(self, y_preds: List[Any], y_trues: List[Any]) -> float:
+        pass
+```
+
+**Sample Custom Metric Adapter Initialization**
+Let's create a Custom metric adapter that parses the inference output for the answer between `<answer> </answer>` tags and then performs an exact match metric.
+
+```python
+from amzn_nova_prompt_optimizer.core.input_adapters.metric_adapter import MetricAdapter
+from typing import List, Any, Dict
+import re
+import json
+
+class CustomMetric(MetricAdapter):
+    def _parse_answer(self, model_output):
+        # Parse Answer between tags
+        match = re.search(r"<answer>(.*?)</answer>", model_output)
+        if not match:
+            return "Choice not found"
+        return match.group(1).lower().strip()
+
+    def _calculate_metrics(self, y_pred: Any, y_true: Any) -> Dict:
+        # Peform Exact Match
+        pred = self._parse_answer(y_pred)
+        ground_truth = self._parse_answer(y_true)
+        return int(pred == ground_truth)
+
+    def apply(self, y_pred: Any, y_true: Any):
+        # Apply to one row of the dataset
+        return self._calculate_metrics(y_pred, y_true)
+
+    def batch_apply(self, y_preds: List[Any], y_trues: List[Any]):
+        # Apply to the whole dataset
+        evals = []
+        for y_pred, y_true in zip(y_preds, y_trues):
+            evals.append(self.apply(y_pred, y_true))
+        return sum(evals)/len(evals)
+
+metric_adapter = CustomMetric()
+```
+
+**Core Functions**
+Apply the metric on a prediction and ground_truth one row at a time
+```python
+y_pred = "The question asks ...... <answer>3</answer>"
+y_true = "<answer>3</answer>"
+
+# Apply the metric on a prediction and ground_truth
+score = metric_adapter.apply(y_pred, y_true)
+
+# score = 1
+```
+
+Apply the metric on a list of prediction and ground_truth i.e. for the dataset
+```python
+y_preds = ["The question asks ...... <answer>3</answer>", "The question asks ...... <answer>5</answer>"]
+y_trues = ["<answer>3</answer>", "<answer>4</answer>"]
+
+# Apply the metric on a list of prediction and ground_truth
+aggregeate_score = metric_adapter.batch_apply(y_preds, y_trues)
+
+# aggregeate_score = 0.5
+```
+
+
+### 5. Optimization Adapter
+**Responsibility:** Load Optimizer, Prompt Adapter, and Optionally Dataset Adapter, Metric Adapter, and Inference Adapter. Perform Optimization and ability to create a Prompt Adapter with the Optimized Prompt.
+
+Sample Optimization Initialization
+```python
+from amzn_nova_prompt_optimizer.core.optimizers import NovaPromptOptimizer
+
+nova_prompt_optimizer = NovaPromptOptimizer(prompt_adapter=prompt_adapter, inference_adapter=inference_adapter, dataset_adapter=train_dataset_adapter, metric_adapter=metric_adapter)
+
+optimized_prompt_adapter = nova_prompt_optimizer.optimize(mode="lite")
+```
+
+We can take a look more deeply into the optimizers in the next section.
+
+
+## Optimizers
+
+### NovaPromptOptimizer
+
+NovaPromptOptimizer is a combination of Meta Prompting using the Nova Guide on prompting and DSPy's MIPROv2 Optimizer using Nova Prompting Tips. 
+NovaPromptOptimizer first runs a meta prompter to identify system instructions and user template from the prompt adapter. 
+Then MIPROv2 is run on top of this to optimize system instructions and identify few-shot samples that need to be added. 
+The few shot samples are added as `converse` format so they are added as User/Assistant turns.
+
+**Requirements:** NovaPromptOptimizer requires Prompt Adapter, Dataset Adapter, Metric Adapter and Inference Adapter.
+
+**Optimization Example**
+```python
+from amzn_nova_prompt_optimizer.core.optimizers import NovaPromptOptimizer
+
+nova_prompt_optimizer = NovaPromptOptimizer(prompt_adapter=prompt_adapter, inference_adapter=inference_adapter, dataset_adapter=train_dataset_adapter, metric_adapter=metric_adapter)
+
+optimized_prompt_adapter = nova_prompt_optimizer.optimize(mode="lite")
+```
+NovaPromptOptimizer uses Premier for Meta Prompting and then uses MIPROv2 with 20 candidates and 50 trials with Premier as Prompting model and task model dependent on the mode it's set at.
+You can specify enable_json_fallback=False to disable the behavior that MIPROv2 will [fallback to use JSONAdapter to parse LM model output](https://github.com/stanfordnlp/dspy/blob/main/dspy/adapters/chat_adapter.py#L44-L51). This will force MIPROv2 use structured output (pydantic model) to parse LM output.
+
+You could also define a custom mode and pass your own parameter values to NovaPromptOptimizer
+
+```python
+from amzn_nova_prompt_optimizer.core.optimizers import NovaPromptOptimizer
+
+nova_prompt_optimizer = NovaPromptOptimizer(prompt_adapter=prompt_adapter, inference_adapter=inference_adapter, dataset_adapter=train_dataset_adapter, metric_adapter=metric_adapter)
+
+optimized_prompt_adapter = nova_prompt_optimizer.optimize(mode="custom", custom_params={"task_model_id": "us.amazon.nova-pro-v1:0",
+    "num_candidates": 10,
+    "num_trials": 15,
+    "max_bootstrapped_demos": 5,
+    "max_labeled_demos": 0
+})
+```
+
+Learn More about the Optimizers [here](docs/Optimizers.md)
+
+## Evaluator
+The SDK also provides a way to baseline prompts and provide evaluation scores.
+The evaluator has the `aggregate_score` and `scores` function.
+
+**Initialization Example**
+```python
+from amzn_nova_prompt_optimizer.core.evaluation import Evaluator
+
+evaluator = Evaluator(nova_mp_optimized_prompt_adapter, test_dataset_adapter, metric_adapter, inference_adapter)
+
+nova_mp_score = evaluator.aggregate_score(model_id="us.amazon.nova-lite-v1:0")
+```
+**Core Functions**
+
+Runs Batch evaluation on the dataset using the batch_apply function of the metric
+```python
+# Uses Batch Apply
+aggregate_score = evaluator.aggregate_score(model_id)
+```
+
+Runs evaluation on the dataset a row at a time and returns the eval results as a whole.
+```python
+# Uses Apply metric. Returns a list of scores.
+scores = evaluator.score(model_id)
+```
+
+Save the eval results.
+```python
+# Save the eval results
+evaluator.save("eval_results.jsonl")
+```
+
+
+**Note: You may come across the below warning. This is when prompt variables are missing from the prompt, the inference runner under the evaluator appends them to the end of the prompt for continuity**
+```python
+WARNING amzn_nova_prompt_optimizer.core.inference: Warn: Prompt Variables not found in User Prompt, injecting them at the end of the prompt
+```
+
+## Optimization Recommendations
+1. Provide representative real-world evaluation sets and split them into training and testing sets. Ensure dataset is balanced on output label when splitting train and test sets.
+2. For evaluation sets, the ground truth column should be as close to the inference output as possible. e.g. If the inference output is {"answer": "POSITIVE"} ground truth should also be in the same format {"answer": "POSITIVE"}
+3. For NovaPromptOptimizer, choose the mode (mode= "premier" | ""pro" | "lite" | "micro") based on your Nova Model of choice. By default, we use "pro".
+4. The `apply` function of the evaluation metric should return a numerical value between 0 and 1 for NovaPromptOptimizer or MIPROv2.
+
+## ⚠️ Preview Status
+
+NovaPromptOptimizer is currently in public preview. During this period:
+- SDK functionality might change as we support more use cases.
+- We welcome feedback and contributions
+
+## Acknowledgements
+* Special acknowledgment to [DSPy](https://github.com/stanfordnlp/dspy) – your innovations continue to inspire us.