Merge pull request #1 from guardrails-ai/fit-n-finish

Fit & Finish

.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

Makefile

@@ -1,8 +1,12 @@
+dev:
+	pip install -e ".[dev]"
+	python3 validator/post-install.py
+
 lint:
 	ruff check .
 
-tests:
-	pytest ./test
+test:
+	pytest ./tests
 
 type:
 	pyright validator

README.md

@@ -30,21 +30,15 @@ In this example, we apply the validator to a string output generated by an LLM.
 
 ```python
 # Import Guard and Validator
+from guardrails import Guard
 from guardrails.hub import CompetitorCheck
 
-# Initialize Validator
-val = CompetitorCheck(
-    competitors=["Apple", "Samsung",]
-    on_fail="noop"
-)
 
 # Setup Guard
-guard = Guard.from_string(
-    validators=[val, ...],
-)
+guard = Guard().use(CompetitorCheck, ["Apple", "Samsung"])
 
-guard.parse("The apple doesn't fall far from the tree.")  # Validator passes
-guard.parse("Apple just released a new iPhone.")  # Validator fails
+guard.validate("The apple doesn't fall far from the tree.")  # Validator passes
+guard.validate("Apple just released a new iPhone.")  # Validator fails
 ```
 
 ### Validating JSON output via Python
@@ -53,13 +47,13 @@ In this example, we apply the validator to a string that is a field within a Pyd
 
 ```python
 # Import Guard and Validator
-from pydantic import BaseModel
+from pydantic import BaseModel, Field
 from guardrails.hub import CompetitorCheck
 from guardrails import Guard
 
 # Initialize Validator
 val = CompetitorCheck(
-    competitors=["Apple", "Samsung",]
+    competitors=["Apple", "Samsung"],
     on_fail="noop"
 )
 
@@ -92,14 +86,14 @@ Initializes a new instance of the Validator class.
 
 **Parameters:**
 
-- **`competitors`** _(list(str))_: List of names of competitors to avoid.
+- **`competitors`** *(List[str]):* List of names of competitors to avoid.
 - **`on_fail`** *(str, Callable):* The policy to enact when a validator fails. If `str`, must be one of `reask`, `fix`, `filter`, `refrain`, `noop`, `exception` or `fix_reask`. Otherwise, must be a function that is called when the validator fails.
 
 </ul>
 
 <br/>
 
-**`__call__(self, value, metadata={}) → ValidationOutcome`**
+**`validate(self, value, metadata={}) → ValidationResult`**
 
 <ul>
 

validator/main.py

@@ -1,3 +1,5 @@
+import nltk
+import spacy
 import re
 from typing import Callable, Dict, List, Optional
 
@@ -10,22 +12,6 @@
     register_validator,
 )
 
-try:
-    import nltk  # type: ignore
-except ImportError:
-    nltk = None  # type: ignore
-
-if nltk is not None:
-    try:
-        nltk.data.find("tokenizers/punkt")
-    except LookupError:
-        nltk.download("punkt")
-
-try:
-    import spacy
-except ImportError:
-    spacy = None
-
 
 @register_validator(name="guardrails/competitor_check", data_type="string")
 class CompetitorCheck(Validator):
@@ -47,18 +33,6 @@ def __init__(
         super().__init__(competitors=competitors, on_fail=on_fail)
         self._competitors = competitors
         model = "en_core_web_trf"
-        if spacy is None:
-            raise ImportError(
-                "You must install spacy in order to use the CompetitorCheck validator."
-            )
-
-        if not spacy.util.is_package(model):
-            logger.info(
-                f"Spacy model {model} not installed. "
-                "Download should start now and take a few minutes."
-            )
-            spacy.cli.download(model)  # type: ignore
-
         self.nlp = spacy.load(model)
 
     def exact_match(self, text: str, competitors: List[str]) -> List[str]:
@@ -134,11 +108,6 @@ def validate(self, value: str, metadata=Dict) -> ValidationResult:
             ValidationResult: The validation result.
         """
 
-        if nltk is None:
-            raise ImportError(
-                "`nltk` library is required for `competitors-check` validator. "
-                "Please install it with `poetry add nltk`."
-            )
         sentences = nltk.sent_tokenize(value)
         flagged_sentences = []
         filtered_sentences = []

validator/post-install.py

@@ -1,3 +1,16 @@
-print("post-install starting...")
-print("This is where you would do things like download nltk tokenizers or login to the HuggingFace hub...")
-print("post-install complete!")
+import nltk
+import spacy
+
+try:
+    nltk.data.find("tokenizers/punkt")
+except LookupError:
+    nltk.download("punkt")
+
+
+model = "en_core_web_trf"
+if not spacy.util.is_package(model):
+    print(
+        f"Spacy model {model} not installed. "
+        "Download should start now and take a few minutes."
+    )
+    spacy.cli.download(model)  # type: ignore