Format README to template

.github/workflows/pr_qc.yml

@@ -15,5 +15,5 @@ jobs:
           python-version: 3.11
       - name: Run qa
         run: |
-          pip install ".[dev]"
+          make dev
           make qa

.github/workflows/publish_pypy.yml

@@ -0,0 +1,18 @@
+name: Publish to Guardrails Hub
+
+on:
+  workflow_dispatch:
+  push:
+    # Publish when new releases are tagged.
+    tags:
+      - '*'
+
+jobs:
+  setup:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Build & Deploy
+        uses: guardrails-ai/guardrails/.github/actions/validator_pypi_publish@main
+        with:
+          guardrails_token: ${{ secrets.GR_GUARDRAILS_TOKEN }}
+          validator_id: guardrails/gibberish_text

Makefile

@@ -1,3 +1,7 @@
+dev:
+	pip install ".[dev]"
+	python validator/post-install.py
+
 lint:
 	ruff check .
 

README.md

@@ -4,7 +4,7 @@
 | --- | --- |
 | Date of development | Feb 15, 2024 |
 | Validator type | Format |
-| Blog | - |
+| Blog |  |
 | License | Apache 2 |
 | Input/Output | Output |
 
@@ -13,12 +13,20 @@
 This validator validates the "cleanliness" of the text generated by a language model. It uses a pre-trained model to determine if the text is coherent and not gibberish. The validator can be used to filter out text that is not coherent or does not make sense.
 
 ## Requirements
-- Dependencies: `nltk`, `transformers`, `torch`
+
+* Dependencies:
+    - guardrails-ai>=0.4.0
+    - `nltk`
+    - `transformers`
+    - `torch`
+
+* Foundation model access keys:
+    - HF_TOKEN
 
 ## Installation
 
 ```bash
-guardrails hub install hub://guardrails/gibberish_text
+$ guardrails hub install hub://guardrails/gibberish_text
 ```
 
 ## Usage Examples
@@ -59,7 +67,7 @@ Validation failed for field with errors: The following sentences in your respons
 ```
 **Note: See how only the first 2 sentences within the failing response are considered gibberish.**
 
-## API Reference
+# API Reference
 
 **`__init__(self, threshold=0.5, validation_method='sentence', on_fail="noop")`**
 <ul>
@@ -78,7 +86,7 @@ Initializes a new instance of the Validator class.
 
 <br/>
 
-**`__call__(self, value, metadata={}) -> ValidationResult`**
+**`validate(self, value, metadata={}) -> ValidationResult`**
 
 <ul>
 

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "gibberish_text"
-version = "0.0.1"
+version = "0.0.2"
 description = "A Guardrails AI validator to detect gibberish text."
 authors = [
     {name = "Guardrails AI", email = "[email protected]"}

validator/main.py

@@ -133,3 +133,5 @@ def validate(self, value: str, metadata: Dict[str, Any]) -> ValidationResult:
 
         if self._validation_method == "full":
             return self.validate_full_text(value, metadata)
+
+        raise ValueError("validation_method must be 'sentence' or 'full'.")

validator/post-install.py

@@ -1,4 +1,3 @@
-import nltk
 from transformers import pipeline
 
 # Download NLTK data if not already present