Merge pull request #3 from jld-adriano/cli

Add cli interface

jaiqu/cli.py

@@ -0,0 +1,66 @@
+import json
+import sys
+
+from typer import Option, Typer
+
+from .jaiqu import translate_schema
+
+typer_app = Typer()
+
+
+@typer_app.command()
+def jaiqu(
+    schema_file: str = Option(..., "-s", "--schema", help="Json schema file path"),
+    data_file: str = Option(
+        None,
+        "-d",
+        "--data",
+        help="Json data file path. if not passed will try to read from stdin",
+    ),
+    quiet: bool = Option(False, "-q", "--quiet", help="Quiet mode, only print errors"),
+    key_hints: str = Option(
+        None,
+        "-k",
+        "--key-hints",
+        help="Extra prompt for the ai to help it complete the task",
+    ),
+    max_retries: int = Option(
+        10,
+        "-r",
+        "--max-retries",
+        help="Max number of retries for the ai to complete the task",
+    ),
+):
+    """
+    Validate and translate a json schema to jq filter
+    """
+    with open(schema_file) as f:
+        output_schema = json.load(f)
+    if data_file is None:
+        if sys.stdin.isatty():
+            sys.exit("Error: No data piped to stdin.")
+        else:
+            if not quiet:
+                print("--data not provided, reading from stdin")
+            data_file = sys.stdin.read()
+            input_json = json.loads(data_file)
+    else:
+        with open(data_file) as f:
+            input_json = json.load(f)
+
+    query = translate_schema(
+        output_schema=output_schema,
+        input_json=input_json,
+        key_hints=key_hints,
+        max_retries=max_retries,
+        quiet=quiet
+    )
+    print(query)
+
+
+def main():
+    typer_app()
+
+
+if __name__ == "__main__":
+    main()

jaiqu/jaiqu.py

@@ -6,7 +6,7 @@
 from .helpers import identify_key, create_jq_string, repair_query, dict_to_jq_filter
 
 
-def validate_schema(input_json: dict, output_schema: dict, openai_api_key: str | None = None, key_hints=None) -> tuple[dict, bool]:
+def validate_schema(input_json: dict, output_schema: dict, openai_api_key: str | None = None, key_hints=None, quiet=False) -> tuple[dict, bool]:
     """Validates the schema of the input JSON against the output schema.
     Args:
         input_json (dict): The input JSON parsed into a dictionary.
@@ -20,7 +20,7 @@ def validate_schema(input_json: dict, output_schema: dict, openai_api_key: str |
 
     results = {}
     valid = True
-    with tqdm(total=len(output_schema['properties']), desc="Validating schema") as pbar:
+    with tqdm(total=len(output_schema['properties']), desc="Validating schema", disable=quiet) as pbar:
         for key, value in output_schema['properties'].items():
             pbar.set_postfix_str(f"Key: {key}", refresh=True)
             response_key, response_reasoning = identify_key(key, value, input_json, openai_api_key, key_hints)
@@ -44,7 +44,7 @@ def validate_schema(input_json: dict, output_schema: dict, openai_api_key: str |
         return results, valid
 
 
-def translate_schema(input_json, output_schema, openai_api_key: str | None = None, key_hints=None, max_retries=10) -> str:
+def translate_schema(input_json, output_schema, openai_api_key: str | None = None, key_hints=None, max_retries=10, quiet=False) -> str:
     """
     Translate the input JSON schema into a filtering query using jq.
 
@@ -64,7 +64,7 @@ def translate_schema(input_json, output_schema, openai_api_key: str | None = Non
         RuntimeError: If failed to validate the jq filter after maximum retries.
     """
 
-    schema_properties, is_valid = validate_schema(input_json, output_schema, key_hints=key_hints, openai_api_key=openai_api_key)
+    schema_properties, is_valid = validate_schema(input_json, output_schema, key_hints=key_hints, openai_api_key=openai_api_key, quiet=quiet)
     if not is_valid:
         raise RuntimeError(
             f"The input JSON does not contain the required data to satisfy the output schema: \n\n{json.dumps(schema_properties, indent=2)}")
@@ -73,7 +73,7 @@ def translate_schema(input_json, output_schema, openai_api_key: str | None = Non
 
     filter_query = {}
 
-    with tqdm(total=len(filtered_schema), desc="Translating schema") as pbar, tqdm(total=max_retries, desc="Retry attempts") as pbar_retries:
+    with tqdm(total=len(filtered_schema), desc="Translating schema", disable=quiet) as pbar, tqdm(total=max_retries, desc="Retry attempts", disable=quiet) as pbar_retries:
         for key, value in filtered_schema.items():
             pbar.set_postfix_str(f"Key: {key}", refresh=True)
             jq_string = create_jq_string(input_json, key, value, openai_api_key)
@@ -101,7 +101,7 @@ def translate_schema(input_json, output_schema, openai_api_key: str | None = Non
     complete_filter = dict_to_jq_filter(filter_query)
     # Validate JSON
     tries = 0
-    with tqdm(total=max_retries, desc="Validation attempts") as pbar_validation:
+    with tqdm(total=max_retries, desc="Validation attempts", disable=quiet) as pbar_validation:
         while True:
             try:
                 result = jq.compile(complete_filter).input(input_json).all()[0]

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "jaiqu"
-version = "0.0.4"
+version = "0.0.5"
 authors = [
   { name = "Alex Reibman", email = "[email protected]" },
   { name = "Howard Gil", email = "[email protected]" },
@@ -21,7 +21,8 @@ classifiers = [
 dependencies = [
     "jq==1.6.0",
     "openai~=1.12.0",
-    "jsonschema==4.21.1"
+    "jsonschema==4.21.1",
+    "typer==0.9.0",
 ]
 
 [project.optional-dependencies]
@@ -34,3 +35,9 @@ dev = [
 [project.urls]
 Homepage = "https://github.com/AgentOps-AI/Jaiqu"
 Issues = "https://github.com/AgentOps-AI/Jaiqu/issues"
+
+[project.entry-points.console_scripts]
+jaiqu = "jaiqu.cli:main"
+
+[tool.setuptools]
+packages = { find = { where = ["."], exclude = ["samples"] } }

requirements.txt

@@ -1,3 +1,4 @@
 jq==1.6.0
 openai>=1.12.0,<2.0.0
-jsonschema==4.21.1
+jsonschema==4.21.1
+typer==0.9.0

samples/data.json

@@ -0,0 +1,11 @@
+{
+    "call.id": "123",
+    "datetime": "2022-01-01",
+    "timestamp": 1640995200,
+    "Address": "123 Main St",
+    "user": {
+        "name": "John Doe",
+        "age": 30,
+        "contact": "[email protected]"
+    }
+}

samples/schema.json

@@ -0,0 +1,22 @@
+{
+    "$schema": "http://json-schema.org/draft-07/schema#",
+    "type": "object",
+    "properties": {
+        "id": {
+            "type": ["string", "null"],
+            "description": "A unique identifier for the record."
+        },
+        "date": {
+            "type": "string",
+            "description": "A string describing the date."
+        },
+        "model": {
+            "type": "string",
+            "description": "A text field representing the model used."
+        }
+    },
+    "required": [
+        "id",
+        "date"
+    ]
+}