p2predict 0.9.4__tar.gz → 0.9.6__tar.gz

{p2predict-0.9.4 → p2predict-0.9.6}/LICENSE

@@ -1,6 +1,21 @@
-PolyForm Noncommercial License 1.0.0
+P2Predict License Exception
+Copyright (c) 2023-2026 Ahmed K. Hafsi
+
+P2Predict is licensed under the PolyForm Noncommercial License 1.0.0 (see below). 
+
+ADDITIONAL GRANT OF PERMISSION:
+Notwithstanding the restrictions in the PolyForm Noncommercial License 1.0.0, the Licensor hereby grants any organization (including for-profit corporations) permission to use this software internally for their own direct operations, procurement, and benchmarking purposes at no cost. 
+
+A commercial license is ONLY required if you:
+1. Provide consulting, advisory, or procurement services to third-party clients using this software.
+2. Embed or integrate this software into a paid commercial product or service.
 
-Copyright (c) 2023 Ahmed K. Hafsi
+For commercial licensing inquiries, please visit p2predict.com.
+
+----------------------------------------------------------------------
+
+PolyForm Noncommercial License 1.0.0
+Copyright (c) 2023-2026 Ahmed K. Hafsi
 
 https://polyformproject.org/licenses/noncommercial/1.0.0
 

{p2predict-0.9.4/src/p2predict.egg-info → p2predict-0.9.6}/PKG-INFO

@@ -1,9 +1,9 @@
 Metadata-Version: 2.4
 Name: p2predict
-Version: 0.9.4
+Version: 0.9.6
 Summary: Parametric price benchmarking for engineering and procurement trade-offs. Agentic first, runs on your data, stays on your machine.
 Author: Ahmed Khalil Hafsi
-License-Expression: PolyForm-Noncommercial-1.0.0
+License-Expression: LicenseRef-PolyForm-Noncommercial-1.0.0-with-Internal-Use-Exception
 Project-URL: Homepage, https://github.com/ahmed-khalil-hafsi/P2Predict
 Project-URL: Repository, https://github.com/ahmed-khalil-hafsi/P2Predict
 Project-URL: Issues, https://github.com/ahmed-khalil-hafsi/P2Predict/issues
@@ -39,7 +39,7 @@ Requires-Dist: halo>=0.0.31
 Requires-Dist: questionary>=1.10
 Requires-Dist: shap<0.50,>=0.44
 Provides-Extra: mcp
-Requires-Dist: mcp>=1.0; extra == "mcp"
+Requires-Dist: mcp<2,>=1.0; extra == "mcp"
 Provides-Extra: dev
 Requires-Dist: pytest>=7; extra == "dev"
 Dynamic: license-file
@@ -55,7 +55,7 @@ Dynamic: license-file
 ![Works with any AI agent](https://img.shields.io/badge/works%20with-any%20AI%20agent-6E56CF) &nbsp;
 ![Local first](https://img.shields.io/badge/local--first-your%20data%20never%20leaves-1F8A4C) &nbsp;
 ![Built for procurement](https://img.shields.io/badge/built%20for-procurement-0B6BCB) &nbsp;
-![License](https://img.shields.io/badge/license-PolyForm%20Noncommercial-555)
+![License](https://img.shields.io/badge/free%20for%20internal%20company%20use-1F8A4C)
 
 ![P2Predict MCP Demo](./documentation/p2predict_mcp_demo.gif)
 
@@ -212,7 +212,7 @@ Each one leads with results, shows where to trust them, and points to exactly wh
 ## FAQ
 
 **Is it really free?**
-Yes, for internal use. P2Predict is source-available under the PolyForm Noncommercial License. No seats, no trial period, no sales call. A commercial license is only needed if you deploy it for clients or embed it in a paid product.
+Yes — including for for-profit companies. Any organization can use P2Predict internally for its own operations, procurement, and benchmarking at no cost. No seats, no trial period, no sales call. A commercial license is only needed if you use it to serve third-party clients (consulting or advisory work) or embed it in a paid product or service.
 
 **How reliable are the numbers?**
 Every estimate comes with a range, not just a number, and the range is honest about how thin your data is: wide where history is sparse, tight where it's deep. On the public case studies, predictions land within roughly 16% of the actual price half the time, and within 73% nine times in ten. That's the model showing its work, not asking you to trust it blind.
@@ -259,10 +259,12 @@ P2Predict comes out of that work: the tools a procurement team actually needs to
 
 ## Licensing
 
-Source-available under the [PolyForm Noncommercial License 1.0.0](LICENSE).
+Source-available under the [PolyForm Noncommercial License 1.0.0, with an additional grant of permission for internal company use](LICENSE).
 
-- **Free for internal use:** use P2Predict inside your own organization at no cost.
-- **Commercial use requires a license:** deploying for clients, embedding it in a paid service, or consulting engagements.
+- **Free for internal use, including for-profit companies.** Any organization may use P2Predict internally for its own operations, procurement, and benchmarking at no cost. You do not need to buy anything, ask permission, or talk to anyone.
+- **A commercial license is only required if you** (1) provide consulting, advisory, or procurement services to third-party clients using P2Predict, or (2) embed or integrate it into a paid product or service.
+
+Despite the "Noncommercial" in the underlying license name, commercial *companies* are explicitly covered by the additional grant — the restriction is on selling P2Predict's use to others, not on being a business.
 
 For a commercial license, a partnership, or to share a procurement dataset, reach out: **[ahmedhafsi.com/contact](https://ahmedhafsi.com/contact/)**.
 

{p2predict-0.9.4 → p2predict-0.9.6}/README.md

@@ -9,7 +9,7 @@
 ![Works with any AI agent](https://img.shields.io/badge/works%20with-any%20AI%20agent-6E56CF)  
 ![Local first](https://img.shields.io/badge/local--first-your%20data%20never%20leaves-1F8A4C)  
 ![Built for procurement](https://img.shields.io/badge/built%20for-procurement-0B6BCB)  
-![License](https://img.shields.io/badge/license-PolyForm%20Noncommercial-555)
+![License](https://img.shields.io/badge/free%20for%20internal%20company%20use-1F8A4C)
 
 ![P2Predict MCP Demo](./documentation/p2predict_mcp_demo.gif)
 
@@ -166,7 +166,7 @@ Each one leads with results, shows where to trust them, and points to exactly wh
 ## FAQ
 
 **Is it really free?**
-Yes, for internal use. P2Predict is source-available under the PolyForm Noncommercial License. No seats, no trial period, no sales call. A commercial license is only needed if you deploy it for clients or embed it in a paid product.
+Yes — including for for-profit companies. Any organization can use P2Predict internally for its own operations, procurement, and benchmarking at no cost. No seats, no trial period, no sales call. A commercial license is only needed if you use it to serve third-party clients (consulting or advisory work) or embed it in a paid product or service.
 
 **How reliable are the numbers?**
 Every estimate comes with a range, not just a number, and the range is honest about how thin your data is: wide where history is sparse, tight where it's deep. On the public case studies, predictions land within roughly 16% of the actual price half the time, and within 73% nine times in ten. That's the model showing its work, not asking you to trust it blind.
@@ -213,10 +213,12 @@ P2Predict comes out of that work: the tools a procurement team actually needs to
 
 ## Licensing
 
-Source-available under the [PolyForm Noncommercial License 1.0.0](LICENSE).
+Source-available under the [PolyForm Noncommercial License 1.0.0, with an additional grant of permission for internal company use](LICENSE).
 
-- **Free for internal use:** use P2Predict inside your own organization at no cost.
-- **Commercial use requires a license:** deploying for clients, embedding it in a paid service, or consulting engagements.
+- **Free for internal use, including for-profit companies.** Any organization may use P2Predict internally for its own operations, procurement, and benchmarking at no cost. You do not need to buy anything, ask permission, or talk to anyone.
+- **A commercial license is only required if you** (1) provide consulting, advisory, or procurement services to third-party clients using P2Predict, or (2) embed or integrate it into a paid product or service.
+
+Despite the "Noncommercial" in the underlying license name, commercial *companies* are explicitly covered by the additional grant — the restriction is on selling P2Predict's use to others, not on being a business.
 
 For a commercial license, a partnership, or to share a procurement dataset, reach out: **[ahmedhafsi.com/contact](https://ahmedhafsi.com/contact/)**.
 

{p2predict-0.9.4 → p2predict-0.9.6}/pyproject.toml

@@ -4,10 +4,10 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "p2predict"
-version = "0.9.4"
+version = "0.9.6"
 description = "Parametric price benchmarking for engineering and procurement trade-offs. Agentic first, runs on your data, stays on your machine."
 readme = "README.md"
-license = "PolyForm-Noncommercial-1.0.0"
+license = "LicenseRef-PolyForm-Noncommercial-1.0.0-with-Internal-Use-Exception"
 license-files = ["LICENSE"]
 authors = [{name = "Ahmed Khalil Hafsi"}]
 requires-python = ">=3.10"
@@ -55,7 +55,7 @@ dependencies = [
 
 [project.optional-dependencies]
 mcp = [
-    "mcp>=1.0",
+    "mcp>=1.0,<2",
 ]
 dev = [
     "pytest>=7",

{p2predict-0.9.4 → p2predict-0.9.6}/src/p2predict/cli/predict.py

@@ -412,6 +412,7 @@ def _print_whatif(console, result: WhatIfResult, target_name: str) -> None:
 
 
 @click.command()
+@click.version_option(package_name="p2predict", message="%(prog)s %(version)s")
 @click.option("-m", "--model", type=click.Path(exists=True),
               help="Path to the trained model file (.model)")
 @click.option("-p", "--predict_using",

{p2predict-0.9.4 → p2predict-0.9.6}/src/p2predict/cli/train.py

@@ -95,6 +95,7 @@ def _feature_outlier_summary_block(summary: dict) -> dict:
 
 
 @click.command()
+@click.version_option(package_name="p2predict", message="%(prog)s %(version)s")
 @click.option("-i", "--input", type=click.Path(exists=True), default=None,
               help="Path to the CSV file containing the training dataset.")
 @click.option("-t", "--target",

{p2predict-0.9.4 → p2predict-0.9.6}/src/p2predict/mcp/server.py

@@ -8,6 +8,8 @@ import argparse
 import asyncio
 import datetime
 import json
+import os
+import sys
 from pathlib import Path
 from typing import Any
 
@@ -1230,6 +1232,73 @@ async def model_resource(model_id: str) -> str:
 # ---------------------------------------------------------------------------
 
 
+def _launch_command() -> list[str]:
+    """The command an MCP client should use to start this server.
+
+    MCP clients do not inherit the shell's PATH, so a bare ``p2predict-mcp``
+    usually fails in the client even when it works in the terminal. Resolve an
+    absolute path instead: prefer the console script sitting next to the
+    running interpreter (the venv layout INSTALL.md builds), and fall back to
+    ``<python> -m p2predict.mcp``, which works even where the console script
+    was never placed on disk or never landed on PATH.
+    """
+    script = Path(sys.executable).parent / (
+        "p2predict-mcp.exe" if os.name == "nt" else "p2predict-mcp"
+    )
+    if script.exists():
+        return [str(script)]
+    # Deliberately NOT resolve()d: sys.executable is already absolute, and
+    # resolving it follows the venv's python symlink out to the base
+    # interpreter — which does not have p2predict on its path, so the emitted
+    # config would fail with ModuleNotFoundError.
+    return [sys.executable, "-m", "p2predict.mcp"]
+
+
+def _client_config_path() -> str:
+    """Where Claude Desktop keeps its config, per platform."""
+    if sys.platform == "darwin":
+        return "~/Library/Application Support/Claude/claude_desktop_config.json"
+    if os.name == "nt":
+        return r"%APPDATA%\Claude\claude_desktop_config.json"
+    return "~/.config/Claude/claude_desktop_config.json"
+
+
+def _print_config(models_dir: Path) -> None:
+    """Print a ready-to-paste MCP client config for this exact install.
+
+    Hand-writing this block is the most error-prone step of setup: the path
+    has to be absolute, and on Windows every backslash has to be doubled.
+    ``json.dumps`` does both correctly, so the user never escapes anything.
+    """
+    command, *extra_args = _launch_command()
+    blob = {
+        "mcpServers": {
+            "p2predict": {
+                "command": command,
+                "args": [*extra_args, "--models-dir", str(models_dir)],
+            }
+        }
+    }
+
+    print("Add this to your Claude Desktop config file:")
+    print(f"  {_client_config_path()}")
+    print()
+    print(json.dumps(blob, indent=2))
+    print()
+    if not models_dir.exists():
+        print(
+            f"Note: {models_dir} doesn't exist yet. That's fine — it's created "
+            "when you train your first model."
+        )
+        print()
+    print(
+        "If the file already has other servers in it, add only the "
+        '"p2predict" block, inside the existing "mcpServers".'
+    )
+    quit_hint = "completely (Cmd+Q)" if sys.platform == "darwin" else "completely"
+    print(f"Then quit Claude Desktop {quit_hint} and reopen it.")
+
+
 def main():
     parser = argparse.ArgumentParser(
         description="P2Predict MCP server — parametric price benchmarking for AI agents"
@@ -1239,15 +1308,33 @@ def main():
         default="models",
         help="Directory containing .model files (default: models)",
     )
+    parser.add_argument(
+        "--version",
+        action="version",
+        version=f"p2predict-mcp {_SERVER_BUILD['version']}",
+    )
+    parser.add_argument(
+        "--print-config",
+        action="store_true",
+        help=(
+            "Print a ready-to-paste Claude Desktop config for this install "
+            "(absolute paths, correctly escaped) and exit."
+        ),
+    )
     args = parser.parse_args()
 
+    models_dir = Path(args.models_dir).resolve()
+
+    if args.print_config:
+        _print_config(models_dir)
+        return
+
     global _registry
-    _registry = ModelRegistry(Path(args.models_dir).resolve())
+    _registry = ModelRegistry(models_dir)
 
     # stderr is safe on stdio transport (stdout is the MCP protocol channel) and
     # shows up in the client's server logs — a quick way to confirm which build
     # is actually running.
-    import sys
     print(
         f"P2Predict MCP server v{_SERVER_BUILD['version']} "
         f"({_SERVER_BUILD['git_sha'] or 'no-git'}) loaded from "

{p2predict-0.9.4 → p2predict-0.9.6/src/p2predict.egg-info}/PKG-INFO

@@ -1,9 +1,9 @@
 Metadata-Version: 2.4
 Name: p2predict
-Version: 0.9.4
+Version: 0.9.6
 Summary: Parametric price benchmarking for engineering and procurement trade-offs. Agentic first, runs on your data, stays on your machine.
 Author: Ahmed Khalil Hafsi
-License-Expression: PolyForm-Noncommercial-1.0.0
+License-Expression: LicenseRef-PolyForm-Noncommercial-1.0.0-with-Internal-Use-Exception
 Project-URL: Homepage, https://github.com/ahmed-khalil-hafsi/P2Predict
 Project-URL: Repository, https://github.com/ahmed-khalil-hafsi/P2Predict
 Project-URL: Issues, https://github.com/ahmed-khalil-hafsi/P2Predict/issues
@@ -39,7 +39,7 @@ Requires-Dist: halo>=0.0.31
 Requires-Dist: questionary>=1.10
 Requires-Dist: shap<0.50,>=0.44
 Provides-Extra: mcp
-Requires-Dist: mcp>=1.0; extra == "mcp"
+Requires-Dist: mcp<2,>=1.0; extra == "mcp"
 Provides-Extra: dev
 Requires-Dist: pytest>=7; extra == "dev"
 Dynamic: license-file
@@ -55,7 +55,7 @@ Dynamic: license-file
 ![Works with any AI agent](https://img.shields.io/badge/works%20with-any%20AI%20agent-6E56CF) &nbsp;
 ![Local first](https://img.shields.io/badge/local--first-your%20data%20never%20leaves-1F8A4C) &nbsp;
 ![Built for procurement](https://img.shields.io/badge/built%20for-procurement-0B6BCB) &nbsp;
-![License](https://img.shields.io/badge/license-PolyForm%20Noncommercial-555)
+![License](https://img.shields.io/badge/free%20for%20internal%20company%20use-1F8A4C)
 
 ![P2Predict MCP Demo](./documentation/p2predict_mcp_demo.gif)
 
@@ -212,7 +212,7 @@ Each one leads with results, shows where to trust them, and points to exactly wh
 ## FAQ
 
 **Is it really free?**
-Yes, for internal use. P2Predict is source-available under the PolyForm Noncommercial License. No seats, no trial period, no sales call. A commercial license is only needed if you deploy it for clients or embed it in a paid product.
+Yes — including for for-profit companies. Any organization can use P2Predict internally for its own operations, procurement, and benchmarking at no cost. No seats, no trial period, no sales call. A commercial license is only needed if you use it to serve third-party clients (consulting or advisory work) or embed it in a paid product or service.
 
 **How reliable are the numbers?**
 Every estimate comes with a range, not just a number, and the range is honest about how thin your data is: wide where history is sparse, tight where it's deep. On the public case studies, predictions land within roughly 16% of the actual price half the time, and within 73% nine times in ten. That's the model showing its work, not asking you to trust it blind.
@@ -259,10 +259,12 @@ P2Predict comes out of that work: the tools a procurement team actually needs to
 
 ## Licensing
 
-Source-available under the [PolyForm Noncommercial License 1.0.0](LICENSE).
+Source-available under the [PolyForm Noncommercial License 1.0.0, with an additional grant of permission for internal company use](LICENSE).
 
-- **Free for internal use:** use P2Predict inside your own organization at no cost.
-- **Commercial use requires a license:** deploying for clients, embedding it in a paid service, or consulting engagements.
+- **Free for internal use, including for-profit companies.** Any organization may use P2Predict internally for its own operations, procurement, and benchmarking at no cost. You do not need to buy anything, ask permission, or talk to anyone.
+- **A commercial license is only required if you** (1) provide consulting, advisory, or procurement services to third-party clients using P2Predict, or (2) embed or integrate it into a paid product or service.
+
+Despite the "Noncommercial" in the underlying license name, commercial *companies* are explicitly covered by the additional grant — the restriction is on selling P2Predict's use to others, not on being a business.
 
 For a commercial license, a partnership, or to share a procurement dataset, reach out: **[ahmedhafsi.com/contact](https://ahmedhafsi.com/contact/)**.
 

{p2predict-0.9.4 → p2predict-0.9.6}/src/p2predict.egg-info/requires.txt

@@ -17,4 +17,4 @@ shap<0.50,>=0.44
 pytest>=7
 
 [mcp]
-mcp>=1.0
+mcp<2,>=1.0

{p2predict-0.9.4 → p2predict-0.9.6}/tests/test_mcp.py

@@ -2,6 +2,7 @@
 from __future__ import annotations
 
 import json
+import os
 from pathlib import Path
 
 import numpy as np
@@ -603,3 +604,75 @@ def test_registry_scan(registry, model_id):
 def test_registry_model_not_found(registry):
     with pytest.raises(FileNotFoundError):
         registry.load("nonexistent_model")
+
+
+# ---------------------------------------------------------------------------
+# --print-config
+# ---------------------------------------------------------------------------
+
+
+def _emitted_config(capsys, models_dir: Path) -> dict:
+    """Run _print_config and parse the JSON block out of what it printed."""
+    mcp_server._print_config(models_dir)
+    text = capsys.readouterr().out
+    return json.loads(text[text.index("{"):text.rindex("}") + 1])
+
+
+def test_print_config_emits_valid_pasteable_json(capsys, tmp_path):
+    """The whole point: the user never hand-escapes anything."""
+    models = tmp_path / "models"
+    models.mkdir()
+    blob = _emitted_config(capsys, models)
+
+    entry = blob["mcpServers"]["p2predict"]
+    assert Path(entry["command"]).is_absolute()
+    assert entry["args"][-2] == "--models-dir"
+    assert entry["args"][-1] == str(models)
+
+
+def test_print_config_uses_absolute_command_not_bare_name(capsys, tmp_path):
+    """MCP clients don't inherit PATH, so a bare command name would fail."""
+    blob = _emitted_config(capsys, tmp_path / "models")
+    command = blob["mcpServers"]["p2predict"]["command"]
+    assert command != "p2predict-mcp"
+    assert Path(command).is_absolute()
+
+
+def test_print_config_flags_missing_models_dir(capsys, tmp_path):
+    mcp_server._print_config(tmp_path / "not_there")
+    assert "doesn't exist yet" in capsys.readouterr().out
+
+
+def test_print_config_stays_quiet_when_models_dir_exists(capsys, tmp_path):
+    models = tmp_path / "models"
+    models.mkdir()
+    mcp_server._print_config(models)
+    assert "doesn't exist yet" not in capsys.readouterr().out
+
+
+def test_launch_command_prefers_console_script(tmp_path, monkeypatch):
+    fake_bin = tmp_path / "bin"
+    fake_bin.mkdir()
+    name = "p2predict-mcp.exe" if os.name == "nt" else "p2predict-mcp"
+    script = fake_bin / name
+    script.touch()
+    monkeypatch.setattr(mcp_server.sys, "executable", str(fake_bin / "python"))
+
+    assert mcp_server._launch_command() == [str(script)]
+
+
+def test_launch_command_falls_back_to_unresolved_interpreter(tmp_path, monkeypatch):
+    """Regression: resolve() followed a venv's python symlink out to the base
+    interpreter, which has no p2predict on its path — the emitted config then
+    died with ModuleNotFoundError. sys.executable must be used as-is."""
+    venv_bin = tmp_path / "venv" / "bin"
+    venv_bin.mkdir(parents=True)
+    real_python = tmp_path / "real_python"
+    real_python.touch()
+    venv_python = venv_bin / "python"
+    venv_python.symlink_to(real_python)
+    monkeypatch.setattr(mcp_server.sys, "executable", str(venv_python))
+
+    command = mcp_server._launch_command()
+    assert command == [str(venv_python), "-m", "p2predict.mcp"]
+    assert command[0] != str(real_python)