proscenium 0.0.1__tar.gz → 0.0.2__tar.gz

{proscenium-0.0.1 → proscenium-0.0.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: proscenium
-Version: 0.0.1
+Version: 0.0.2
 Summary: Frame AI Agents
 License: ASFv2
 Author: Adam Pingel
@@ -26,7 +26,6 @@ Requires-Dist: pymilvus_model (>=0.3.1,<0.4.0)
 Requires-Dist: python-dotenv (>=1.0.1,<2.0.0)
 Requires-Dist: rich (>=13.9.4,<14.0.0)
 Requires-Dist: stringcase (>=1.2.0,<2.0.0)
-Requires-Dist: thespian (>=4.0.0,<5.0.0)
 Requires-Dist: tiktoken (>=0.9.0,<0.10.0)
 Requires-Dist: typer (>=0.15.2,<0.16.0)
 Description-Content-Type: text/markdown
@@ -34,17 +33,14 @@ Description-Content-Type: text/markdown
 # Proscenium
 
 [![CI](https://github.com/The-AI-Alliance/proscenium/actions/workflows/pytest.yml/badge.svg)](https://github.com/The-AI-Alliance/proscenium/actions/workflows/pytest.yml)
+[![PyPI](https://img.shields.io/pypi/v/proscenium)](https://pypi.org/project/proscenium/)
 [![License](https://img.shields.io/github/license/The-AI-Alliance/proscenium)](https://github.com/The-AI-Alliance/proscenium/tree/main?tab=Apache-2.0-1-ov-file#readme)
 [![Issues](https://img.shields.io/github/issues/The-AI-Alliance/proscenium)](https://github.com/The-AI-Alliance/proscenium/issues)
 [![GitHub stars](https://img.shields.io/github/stars/The-AI-Alliance/proscenium?style=social)](https://github.com/The-AI-Alliance/proscenium/stargazers)
 
-Proscenium is a small, experimental library of composable glue that allows for
-succinct construction of enterprise AI applications.  It is in early development.
+Proscenium is a small, experimental library of composable glue that allows for succinct construction of enterprise AI applications. It was started in February 2025 and is still in early development.
 
-It is also set of simple demonstration applications that elucidate aspects of application and library design.
-
-Currently, proscenium development prioritizes support for domains where
-the creation and use of structured data is critical.
+Currently, proscenium development prioritizes support for domains where the creation and use of structured data is critical.
 
 See the [website](https://the-ai-alliance.github.io/proscenium/) for quickstart info, goals, and other links.
 

{proscenium-0.0.1 → proscenium-0.0.2}/README.md

@@ -1,17 +1,14 @@
 # Proscenium
 
 [![CI](https://github.com/The-AI-Alliance/proscenium/actions/workflows/pytest.yml/badge.svg)](https://github.com/The-AI-Alliance/proscenium/actions/workflows/pytest.yml)
+[![PyPI](https://img.shields.io/pypi/v/proscenium)](https://pypi.org/project/proscenium/)
 [![License](https://img.shields.io/github/license/The-AI-Alliance/proscenium)](https://github.com/The-AI-Alliance/proscenium/tree/main?tab=Apache-2.0-1-ov-file#readme)
 [![Issues](https://img.shields.io/github/issues/The-AI-Alliance/proscenium)](https://github.com/The-AI-Alliance/proscenium/issues)
 [![GitHub stars](https://img.shields.io/github/stars/The-AI-Alliance/proscenium?style=social)](https://github.com/The-AI-Alliance/proscenium/stargazers)
 
-Proscenium is a small, experimental library of composable glue that allows for
-succinct construction of enterprise AI applications.  It is in early development.
+Proscenium is a small, experimental library of composable glue that allows for succinct construction of enterprise AI applications. It was started in February 2025 and is still in early development.
 
-It is also set of simple demonstration applications that elucidate aspects of application and library design.
-
-Currently, proscenium development prioritizes support for domains where
-the creation and use of structured data is critical.
+Currently, proscenium development prioritizes support for domains where the creation and use of structured data is critical.
 
 See the [website](https://the-ai-alliance.github.io/proscenium/) for quickstart info, goals, and other links.
 

proscenium-0.0.2/proscenium/scripts/chunk_space.py

@@ -0,0 +1,39 @@
+import logging
+
+from pymilvus import MilvusClient
+from pymilvus import model
+
+from proscenium.verbs.read import load_file
+from proscenium.verbs.chunk import documents_to_chunks_by_characters
+from proscenium.verbs.vector_database import create_collection
+from proscenium.verbs.vector_database import add_chunks_to_vector_db
+from proscenium.verbs.display.milvus import collection_panel
+
+
+def make_vector_db_builder(
+    data_files: list[str],
+    vector_db_client: MilvusClient,
+    embedding_fn: model.dense.SentenceTransformerEmbeddingFunction,
+    collection_name: str,
+):
+
+    def build():
+
+        create_collection(
+            vector_db_client, embedding_fn, collection_name, overwrite=True
+        )
+
+        for data_file in data_files:
+
+            documents = load_file(data_file)
+            chunks = documents_to_chunks_by_characters(documents)
+            logging.info("Data file %s has %s chunks", data_file, len(chunks))
+
+            info = add_chunks_to_vector_db(
+                vector_db_client, embedding_fn, chunks, collection_name
+            )
+            logging.info("%s chunks inserted", info["insert_count"])
+
+        logging.info(collection_panel(vector_db_client, collection_name))
+
+    return build

{proscenium-0.0.1 → proscenium-0.0.2}/proscenium/scripts/document_enricher.py

@@ -1,12 +1,13 @@
 from typing import List
 from typing import Callable
-from typing import Any
+from typing import Optional
 
 import time
+import logging
 from pydantic import BaseModel
 
-from rich import print
 from rich.panel import Panel
+from rich.console import Console
 from rich.progress import Progress
 
 from langchain_core.documents.base import Document
@@ -22,11 +23,12 @@ def extract_from_document_chunks(
     chunk_extraction_template: str,
     chunk_extract_clazz: type[BaseModel],
     delay: float,
-    verbose: bool = False,
+    console: Optional[Console] = None,
 ) -> List[BaseModel]:
 
-    print(doc_as_rich(doc))
-    print()
+    if console is not None:
+        console.print(doc_as_rich(doc))
+        console.print()
 
     extract_models = []
 
@@ -40,9 +42,9 @@ def extract_from_document_chunks(
             chunk.page_content,
         )
 
-        if verbose:
-            print("Extract model in chunk", i + 1, "of", len(chunks))
-            print(Panel(str(ce)))
+        logging.info("Extract model in chunk %s of %s", i + 1, len(chunks))
+        if console is not None:
+            console.print(Panel(str(ce)))
 
         extract_models.append(ce)
         time.sleep(delay)
@@ -55,7 +57,7 @@ def enrich_documents(
     extract_from_doc_chunks: Callable[[Document], List[BaseModel]],
     doc_enrichments: Callable[[Document, list[BaseModel]], BaseModel],
     enrichments_jsonl_file: str,
-    verbose: bool = False,
+    console: Optional[Console] = None,
 ) -> None:
 
     docs = retrieve_documents()
@@ -70,11 +72,11 @@ def enrich_documents(
 
             for doc in docs:
 
-                chunk_extract_models = extract_from_doc_chunks(doc, verbose)
+                chunk_extract_models = extract_from_doc_chunks(doc)
                 enrichments = doc_enrichments(doc, chunk_extract_models)
                 enrichments_json = enrichments.model_dump_json()
                 f.write(enrichments_json + "\n")
 
                 progress.update(task_enrich, advance=1)
 
-        print("Wrote document enrichments to", enrichments_jsonl_file)
+        logging.info("Wrote document enrichments to %s", enrichments_jsonl_file)

{proscenium-0.0.1 → proscenium-0.0.2}/proscenium/scripts/entity_resolver.py

@@ -1,6 +1,7 @@
 from typing import Optional
-from rich import print
+import logging
 
+from rich.console import Console
 from langchain_core.documents.base import Document
 from neo4j import Driver
 
@@ -21,27 +22,27 @@ class Resolver:
         cypher: str,
         field_name: str,
         collection_name: str,
-        embedding_model_id: str,
     ):
         self.cypher = cypher
         self.field_name = field_name
         self.collection_name = collection_name
-        self.embedding_model_id = embedding_model_id
 
 
 def load_entity_resolver(
     driver: Driver,
     resolvers: list[Resolver],
+    embedding_model_id: str,
     milvus_uri: str,
+    console: Optional[Console] = None,
 ) -> None:
 
     vector_db_client = vector_db(milvus_uri, overwrite=True)
-    print("Vector db stored at", milvus_uri)
+    logging.info("Vector db stored at %s", milvus_uri)
 
     for resolver in resolvers:
 
-        embedding_fn = embedding_function(resolver.embedding_model_id)
-        print("Embedding model", resolver.embedding_model_id)
+        embedding_fn = embedding_function(embedding_model_id)
+        logging.info("Embedding model %s", embedding_model_id)
 
         values = []
         with driver.session() as session:
@@ -49,15 +50,18 @@ def load_entity_resolver(
             new_values = [Document(record[resolver.field_name]) for record in result]
             values.extend(new_values)
 
-        print("Loading entity resolver into vector db", resolver.collection_name)
+        logging.info(
+            "Loading entity resolver into vector db %s", resolver.collection_name
+        )
         create_collection(
             vector_db_client, embedding_fn, resolver.collection_name, overwrite=True
         )
         info = add_chunks_to_vector_db(
             vector_db_client, embedding_fn, values, resolver.collection_name
         )
-        print(info["insert_count"], "chunks inserted")
-        print(collection_panel(vector_db_client, resolver.collection_name))
+        logging.info("%s chunks inserted ", info["insert_count"])
+        if console is not None:
+            console.print(collection_panel(vector_db_client, resolver.collection_name))
 
     vector_db_client.close()
 
@@ -68,10 +72,12 @@ def find_matching_objects(
     resolver: Resolver,
 ) -> Optional[str]:
 
-    print("Loading collection", resolver.collection_name)
+    logging.info("Loading collection", resolver.collection_name)
     vector_db_client.load_collection(resolver.collection_name)
 
-    print("Finding entity matches for", approximate, "using", resolver.collection_name)
+    logging.info(
+        "Finding entity matches for", approximate, "using", resolver.collection_name
+    )
 
     hits = closest_chunks(
         vector_db_client,
@@ -82,8 +88,8 @@ def find_matching_objects(
     )
     # TODO apply distance threshold
     for match in [head["entity"]["text"] for head in hits[:1]]:
-        print("Closest match:", match)
+        logging.info("Closest match:", match)
         return match
 
-    print("No match found")
+    logging.info("No match found")
     return None

proscenium-0.0.2/proscenium/scripts/graph_rag.py

@@ -0,0 +1,56 @@
+from typing import Callable
+from typing import Optional
+
+import logging
+
+from rich.console import Console
+
+from pydantic import BaseModel
+from uuid import uuid4, UUID
+from neo4j import Driver
+
+
+def query_to_prompts(
+    query: str,
+    query_extraction_model_id: str,
+    milvus_uri: str,
+    driver: Driver,
+    query_extract: Callable[
+        [str, str], BaseModel
+    ],  # (query_str, query_extraction_model_id) -> QueryExtractions
+    query_extract_to_graph: Callable[
+        [str, UUID, BaseModel], None
+    ],  # query, query_id, extract
+    query_extract_to_context: Callable[
+        [BaseModel, str, Driver, str, Optional[Console]], BaseModel
+    ],  # (QueryExtractions, query_str, Driver, milvus_uri) -> Context
+    context_to_prompts: Callable[
+        [BaseModel], tuple[str, str]
+    ],  # Context -> (system_prompt, user_prompt)
+    console: Optional[Console] = None,
+) -> str:
+
+    query_id = uuid4()
+
+    logging.info("Extracting information from the question")
+
+    extract = query_extract(query, query_extraction_model_id)
+    if extract is None:
+        logging.info("Unable to extract information from that question")
+        return None
+
+    logging.info("Extract: %s", extract)
+
+    logging.info("Storing the extracted information in the graph")
+    query_extract_to_graph(query, query_id, extract, driver)
+
+    logging.info("Forming context from the extracted information")
+    context = query_extract_to_context(
+        extract, query, driver, milvus_uri, console=console
+    )
+
+    logging.info("Context: %s", context)
+
+    prompts = context_to_prompts(context)
+
+    return prompts

{proscenium-0.0.1 → proscenium-0.0.2}/proscenium/scripts/knowledge_graph.py

@@ -1,10 +1,10 @@
 from typing import Callable
 from typing import Any
 
+import logging
 import json
 from pydantic import BaseModel
 
-from rich import print
 from rich.progress import Progress
 
 from neo4j import Driver
@@ -17,7 +17,7 @@ def load_knowledge_graph(
     doc_enrichments_to_graph: Callable[[Any, BaseModel], None],
 ) -> None:
 
-    print("Parsing enrichments from", enrichments_jsonl_file)
+    logging.info("Parsing enrichments from %s", enrichments_jsonl_file)
 
     enrichmentss = []
     with open(enrichments_jsonl_file, "r") as f:

{proscenium-0.0.1 → proscenium-0.0.2}/proscenium/scripts/rag.py

@@ -1,7 +1,7 @@
-from typing import List, Dict
+from typing import List, Dict, Optional
+import logging
 
-from rich import print
-from rich.panel import Panel
+from rich.console import Console
 
 from pymilvus import MilvusClient
 from pymilvus import model
@@ -44,20 +44,16 @@ def answer_question(
     vector_db_client: MilvusClient,
     embedding_fn: model.dense.SentenceTransformerEmbeddingFunction,
     collection_name: str,
-    verbose: bool = False,
+    console: Optional[Console] = None,
 ) -> str:
 
-    print(Panel(query, title="User"))
-
     chunks = closest_chunks(vector_db_client, embedding_fn, query, collection_name)
-    if verbose:
-        print("Found", len(chunks), "closest chunks")
-        print(chunk_hits_table(chunks))
+    logging.info("Found %s closest chunks", len(chunks))
+    logging.info(chunk_hits_table(chunks))
 
     prompt = rag_prompt(chunks, query)
-    if verbose:
-        print("RAG prompt created. Calling inference at", model_id, "\n\n")
+    logging.info("RAG prompt created. Calling inference at %s", model_id)
 
-    answer = complete_simple(model_id, rag_system_prompt, prompt, rich_output=verbose)
+    answer = complete_simple(model_id, rag_system_prompt, prompt, console=console)
 
     return answer

{proscenium-0.0.1 → proscenium-0.0.2}/proscenium/scripts/tools.py

@@ -1,47 +1,15 @@
-from typing import List
+from typing import Optional
+import logging
 
-from rich import print
+from rich.console import Console
 from rich.panel import Panel
 from rich.text import Text
-from thespian.actors import Actor
-
-from gofannon.base import BaseTool
 
 from proscenium.verbs.complete import (
     complete_for_tool_applications,
     evaluate_tool_calls,
     complete_with_tool_results,
 )
-from proscenium.verbs.invoke import process_tools
-
-
-def tool_applier_actor_class(
-    tools: List[BaseTool],
-    system_message: str,
-    model_id: str,
-    temperature: float = 0.75,
-    rich_output: bool = False,
-):
-
-    tool_map, tool_desc_list = process_tools(tools)
-
-    class ToolApplier(Actor):
-
-        def receiveMessage(self, message, sender):
-
-            response = apply_tools(
-                model_id=model_id,
-                system_message=system_message,
-                message=message,
-                tool_desc_list=tool_desc_list,
-                tool_map=tool_map,
-                temperature=temperature,
-                rich_output=rich_output,
-            )
-
-            self.send(sender, response)
-
-    return ToolApplier
 
 
 def apply_tools(
@@ -51,7 +19,7 @@ def apply_tools(
     tool_desc_list: list,
     tool_map: dict,
     temperature: float = 0.75,
-    rich_output: bool = False,
+    console: Optional[Console] = None,
 ) -> str:
 
     messages = [
@@ -60,35 +28,33 @@ def apply_tools(
     ]
 
     response = complete_for_tool_applications(
-        model_id, messages, tool_desc_list, temperature, rich_output
+        model_id, messages, tool_desc_list, temperature, console
     )
 
     tool_call_message = response.choices[0].message
 
     if tool_call_message.tool_calls is None or len(tool_call_message.tool_calls) == 0:
 
-        if rich_output:
-            print(
+        if console is not None:
+            console.print(
                 Panel(
                     Text(str(tool_call_message.content)),
                     title="Tool Application Response",
                 )
             )
 
-        print("No tool applications detected")
+        logging.info("No tool applications detected")
 
         return tool_call_message.content
 
     else:
 
-        if rich_output:
-            print(
+        if console is not None:
+            console.print(
                 Panel(Text(str(tool_call_message)), title="Tool Application Response")
             )
 
-        tool_evaluation_messages = evaluate_tool_calls(
-            tool_call_message, tool_map, rich_output
-        )
+        tool_evaluation_messages = evaluate_tool_calls(tool_call_message, tool_map)
 
         result = complete_with_tool_results(
             model_id,
@@ -97,7 +63,7 @@ def apply_tools(
             tool_evaluation_messages,
             tool_desc_list,
             temperature,
-            rich_output,
+            console,
         )
 
         return result

{proscenium-0.0.1 → proscenium-0.0.2}/proscenium/verbs/complete.py

@@ -37,10 +37,12 @@ Valid model ids:
 - `ollama:granite3.1-dense:2b`
 """
 
+from typing import Optional
 from typing import Any
-
+import logging
 import json
-from rich import print
+
+from rich.console import Console
 from rich.console import Group
 from rich.panel import Panel
 from rich.table import Table
@@ -63,14 +65,14 @@ def complete_simple(
     model_id: str, system_prompt: str, user_prompt: str, **kwargs
 ) -> str:
 
-    rich_output = kwargs.pop("rich_output", False)
+    console = kwargs.pop("console", None)
 
     messages = [
         {"role": "system", "content": system_prompt},
         {"role": "user", "content": user_prompt},
     ]
 
-    if rich_output:
+    if console is not None:
 
         kwargs_text = "\n".join([str(k) + ": " + str(v) for k, v in kwargs.items()])
 
@@ -90,34 +92,30 @@ model_id: {model_id}
         call_panel = Panel(
             Group(params_text, messages_table), title="complete_simple call"
         )
-        print(call_panel)
+        console.print(call_panel)
 
     response = client.chat.completions.create(
         model=model_id, messages=messages, **kwargs
     )
     response = response.choices[0].message.content
 
-    if rich_output:
-        print(Panel(response, title="Response"))
+    if console is not None:
+        console.print(Panel(response, title="Response"))
 
     return response
 
 
-def evaluate_tool_call(
-    tool_map: dict, tool_call: ChatCompletionMessageToolCall, rich_output: bool = False
-) -> Any:
+def evaluate_tool_call(tool_map: dict, tool_call: ChatCompletionMessageToolCall) -> Any:
 
     function_name = tool_call.function.name
     # TODO validate the arguments?
     function_args = json.loads(tool_call.function.arguments)
 
-    if rich_output:
-        print(f"Evaluating tool call: {function_name} with args {function_args}")
+    logging.info(f"Evaluating tool call: {function_name} with args {function_args}")
 
     function_response = tool_map[function_name](**function_args)
 
-    if rich_output:
-        print(f"   Response: {function_response}")
+    logging.info(f"   Response: {function_response}")
 
     return function_response
 
@@ -134,23 +132,19 @@ def tool_response_message(
     }
 
 
-def evaluate_tool_calls(
-    tool_call_message, tool_map: dict, rich_output: bool = False
-) -> list[dict]:
+def evaluate_tool_calls(tool_call_message, tool_map: dict) -> list[dict]:
 
     tool_call: ChatCompletionMessageToolCall
 
-    if rich_output:
-        print("Evaluating tool calls")
+    logging.info("Evaluating tool calls")
 
     new_messages: list[dict] = []
 
     for tool_call in tool_call_message.tool_calls:
-        function_response = evaluate_tool_call(tool_map, tool_call, rich_output)
+        function_response = evaluate_tool_call(tool_map, tool_call)
         new_messages.append(tool_response_message(tool_call, function_response))
 
-    if rich_output:
-        print("Tool calls evaluated")
+    logging.info("Tool calls evaluated")
 
     return new_messages
 
@@ -160,10 +154,10 @@ def complete_for_tool_applications(
     messages: list,
     tool_desc_list: list,
     temperature: float,
-    rich_output: bool = False,
+    console: Optional[Console] = None,
 ):
 
-    if rich_output:
+    if console is not None:
         panel = complete_with_tools_panel(
             "complete for tool applications",
             model_id,
@@ -171,7 +165,7 @@ def complete_for_tool_applications(
             messages,
             temperature,
         )
-        print(panel)
+        console.print(panel)
 
     response = client.chat.completions.create(
         model=model_id,
@@ -190,13 +184,13 @@ def complete_with_tool_results(
     tool_evaluation_messages: list[dict],
     tool_desc_list: list,
     temperature: float,
-    rich_output: bool = False,
+    console: Optional[Console] = None,
 ):
 
     messages.append(tool_call_message)
     messages.extend(tool_evaluation_messages)
 
-    if rich_output:
+    if console is not None:
         panel = complete_with_tools_panel(
             "complete call with tool results",
             model_id,
@@ -204,7 +198,7 @@ def complete_with_tool_results(
             messages,
             temperature,
         )
-        print(panel)
+        console.print(panel)
 
     response = client.chat.completions.create(
         model=model_id,

{proscenium-0.0.1 → proscenium-0.0.2}/proscenium/verbs/display/__init__.py

@@ -4,6 +4,6 @@ from rich.text import Text
 def header() -> Text:
     text = Text()
     text.append("Proscenium 🎭\n", style="bold")
-    text.append("The AI Alliance\n", style="bold")
+    text.append("https://the-ai-alliance.github.io/proscenium/\n")
     # TODO version, timestamp, ...
     return text

{proscenium-0.0.1 → proscenium-0.0.2}/proscenium/verbs/extract.py

@@ -1,4 +1,6 @@
+from typing import Optional
 import logging
+from rich.console import Console
 from string import Formatter
 
 import json
@@ -36,7 +38,7 @@ def extract_to_pydantic_model(
     extraction_template: str,
     clazz: type[BaseModel],
     text: str,
-    verbose: bool = False,
+    console: Optional[Console] = None,
 ) -> BaseModel:
 
     extract_str = complete_simple(
@@ -47,7 +49,7 @@ def extract_to_pydantic_model(
             "type": "json_object",
             "schema": clazz.model_json_schema(),
         },
-        rich_output=verbose,
+        console=console,
     )
 
     logging.info("complete_to_pydantic_model: extract_str = <<<%s>>>", extract_str)

{proscenium-0.0.1 → proscenium-0.0.2}/proscenium/verbs/read.py

@@ -42,12 +42,10 @@ from pathlib import Path
 async def url_to_file(url: HttpUrl, data_file: Path, overwrite: bool = False):
 
     if data_file.exists() and not overwrite:
-        # print(f"File {data_file} exists. Use overwrite=True to replace.")
         return
 
     async with httpx.AsyncClient() as client:
 
-        # print(f"Downloading {url} to {data_file}...")
         response = await client.get(url)
         response.raise_for_status()
 

{proscenium-0.0.1 → proscenium-0.0.2}/proscenium/verbs/vector_database.py

@@ -1,5 +1,6 @@
-from typing import Dict, List, Optional
+from typing import Dict, List
 
+import logging
 from pathlib import Path
 from langchain_core.documents.base import Document
 from pymilvus import MilvusClient
@@ -54,15 +55,14 @@ def vector_db(
         if file_path.exists():
             if overwrite:
                 file_path.unlink()
-                print("Deleted existing vector db file", file_path)
+                logging.info("Deleted existing vector db file %s", file_path)
             else:
-                print(
-                    "Using existing",
+                logging.info(
+                    "Using existing %s file. Use overwrite=True to replace.",
                     uri_fields[2],
-                    "file. Use overwrite=True to replace.",
                 )
         else:
-            print("Creating new vector db file", file_path)
+            logging.info("Creating new vector db file %s", file_path)
 
         client = MilvusClient(uri=str(file_path))
 
@@ -100,7 +100,7 @@ def create_collection(
     client.create_index(
         collection_name=collection_name, index_params=index_params, sync=True
     )
-    print("Created collection", collection_name)
+    logging.info("Created collection %s", collection_name)
 
 
 def add_chunks_to_vector_db(

{proscenium-0.0.1 → proscenium-0.0.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "proscenium"
-version = "0.0.1"
+version = "0.0.2"
 description = "Frame AI Agents"
 authors = ["Adam Pingel <oss@pingel.org>"]
 license = "ASFv2"
@@ -15,7 +15,6 @@ typer = "^0.15.2"
 python-dotenv = "^1.0.1"
 pydantic = "^2.10.6"
 stringcase = "^1.2.0"
-thespian = "^4.0.0"
 docstring_parser = "^0.16"
 pymilvus = {version = "^2.5.4"}
 pymilvus_model = {version = "^0.3.1"}

proscenium-0.0.1/proscenium/scripts/chunk_space.py

@@ -1,33 +0,0 @@
-from rich import print
-
-from pymilvus import MilvusClient
-from pymilvus import model
-
-from proscenium.verbs.read import load_file
-from proscenium.verbs.chunk import documents_to_chunks_by_characters
-from proscenium.verbs.vector_database import create_collection
-from proscenium.verbs.vector_database import add_chunks_to_vector_db
-from proscenium.verbs.display.milvus import collection_panel
-
-
-def build_vector_db(
-    data_files: list[str],
-    vector_db_client: MilvusClient,
-    embedding_fn: model.dense.SentenceTransformerEmbeddingFunction,
-    collection_name: str,
-):
-
-    create_collection(vector_db_client, embedding_fn, collection_name, overwrite=True)
-
-    for data_file in data_files:
-
-        documents = load_file(data_file)
-        chunks = documents_to_chunks_by_characters(documents)
-        print("Data file", data_file, "has", len(chunks), "chunks")
-
-        info = add_chunks_to_vector_db(
-            vector_db_client, embedding_fn, chunks, collection_name
-        )
-        print(info["insert_count"], "chunks inserted")
-
-    print(collection_panel(vector_db_client, collection_name))

proscenium-0.0.1/proscenium/scripts/graph_rag.py

@@ -1,43 +0,0 @@
-from typing import Callable
-
-from pydantic import BaseModel
-
-from rich import print
-from rich.panel import Panel
-
-from neo4j import Driver
-
-from proscenium.verbs.complete import complete_simple
-
-
-def query_to_prompts(
-    question: str,
-    query_extraction_model_id: str,
-    milvus_uri: str,
-    driver: Driver,
-    query_extract: Callable[
-        [str, str, bool], BaseModel
-    ],  # (query_str, query_extraction_model_id) -> QueryExtractions
-    extract_to_context: Callable[
-        [BaseModel, str, Driver, str, bool], BaseModel
-    ],  # (QueryExtractions, query_str, Driver, milvus_uri) -> Context
-    context_to_prompts: Callable[
-        [BaseModel, bool], tuple[str, str]
-    ],  # Context -> (system_prompt, user_prompt)
-    verbose: bool = False,
-) -> str:
-
-    print("Extracting information from the question")
-    extract = query_extract(question, query_extraction_model_id, verbose)
-    if extract is None:
-        print("Unable to extract information from that question")
-        return None
-    print("Extract:", extract)
-
-    print("Forming context from the extracted information")
-    context = extract_to_context(extract, question, driver, milvus_uri, verbose)
-    print("Context:", context)
-
-    prompts = context_to_prompts(context, verbose)
-
-    return prompts

proscenium-0.0.1/proscenium/verbs/know.py

@@ -1,9 +0,0 @@
-from neo4j import GraphDatabase
-from neo4j import Driver
-
-
-def knowledge_graph_client(uri: str, username: str, password: str) -> Driver:
-
-    driver = GraphDatabase.driver(uri, auth=(username, password))
-
-    return driver