ProjectFlow: default tasks_to_skip to None; retire the _old_spec archive

.vscode/settings.json

@@ -1,16 +1,16 @@
 {
-  "python.testing.pytestEnabled": false,
-  "python.testing.unittestEnabled": true,
+  "python.testing.pytestEnabled": true,
+  "python.testing.unittestEnabled": false,
   "python.testing.pytestArgs": [
     "hazelbean_tests"
   ],
   "python.testing.cwd": "${workspaceFolder}",
   "python.testing.autoTestDiscoverOnSaveEnabled": true,
-  
-  // Conda environment configuration
-  "python.defaultInterpreterPath": "/opt/homebrew/Caskroom/mambaforge/base/envs/hazelbean_env/bin/python",
+
+  // Conda environment configuration (interpreter is machine-specific — set it
+  // via the "Python: Select Interpreter" command, not here)
   "python.condaPath": "/opt/homebrew/bin/conda",
-  
+
   // Terminal activation
   "python.terminal.activateEnvironment": true,
   "python.terminal.activateEnvInCurrentTerminal": true,
@@ -24,8 +24,6 @@
     "htmlcov": true
   },
 
-  // Pytest-specific settings
-  "python.testing.pytestPath": "/opt/homebrew/Caskroom/mambaforge/base/envs/hazelbean_env/bin/pytest",
   "python-envs.defaultEnvManager": "ms-python.python:conda",
   "python-envs.defaultPackageManager": "ms-python.python:conda",
   "python.testing.unittestArgs": [
@@ -34,6 +32,13 @@
     "./hazelbean_tests",
     "-p",
     "test_*.py"
+  ],
+  "python-envs.pythonProjects": [
+    {
+      "path": ".",
+      "envManager": "ms-python.python:conda",
+      "packageManager": "ms-python.python:conda"
+    }
   ]
 }
 

CHANGELOG.md

@@ -1,5 +1,15 @@
 # Hazelbean Changelog
 
+## Unreleased
+
+#### New Features
+
+* `ProjectFlow` initializes `tasks_to_skip` to `None`, so run files on the caller-configured convention can call `p.skip_tasks(p.tasks_to_skip)` without a `getattr` guard.
+
+#### Chores
+
+* Removed `hazelbean/project_flow_old_spec.py`. Git history is the archive for pre-conversion forms; the `_old_spec` twins are being retired stack-wide.
+
 ## v1.8.0 (2026-01-16)
 
 #### New Features

CLAUDE.md

@@ -2,15 +2,23 @@
 
 This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
 
+Hazelbean is the shared base library for the earth-economy devstack. Shared
+devstack guidance (reuse rule + ownership map + EE Spec conventions) is imported
+here so it applies whenever you work in this repo:
+
+@../../earth_economy_devstack/devstack_guidance.md
+
 ## Project Overview
 
 Hazelbean is a geospatial processing library built on GDAL, NumPy, SciPy, Cython, PyGeoProcessing, and TaskGraph. It provides tools for sustainability science, ecosystem service assessment, and land-use modeling.
 
 ## Development Environment
 
-**Required**: Always activate the conda environment before any development work:
+**Required**: Always activate a conda environment with hazelbean's dependencies
+before any development work. The environment name is contributor-specific — use
+your own (see your user-level Claude config):
 ```bash
-conda activate hazelbean_env
+conda activate <your-env>
 ```
 
 **Installation** (after environment activation):

hazelbean/__init__.py

@@ -1,6 +1,21 @@
 from __future__ import division, absolute_import, print_function
 import sys, time, os
 
+
+conda_prefix = os.path.dirname(sys.executable)
+# On Linux/Mac, python is in env/bin/, so go up one more level
+if os.path.basename(conda_prefix) in ("bin", "Scripts"):
+    conda_prefix = os.path.dirname(conda_prefix)
+gdal_bin = os.path.join(conda_prefix, "Library", "bin")
+if os.path.exists(gdal_bin) and gdal_bin not in os.environ["PATH"]:
+    os.environ["PATH"] = gdal_bin + os.pathsep + os.environ["PATH"]
+
+# Per-machine, never-committed settings (e.g. gtappy's solve-backend connection vars)
+# live in ~/.config/hazelbean/machine.env and are loaded here into any env-var gaps.
+# Real environment variables always win. See hazelbean/machine_env.py.
+from hazelbean.machine_env import load_user_machine_env
+load_user_machine_env()
+
 import hazelbean.config # Needs to be imported before core so that hb.config.LAST_TIME_CHECK is set for hb.timer()
 from hazelbean.config import *
 
@@ -15,6 +30,15 @@
 pd.set_option('display.max_columns', 20)
 pd.set_option('display.width', 200)
 pd.set_option('display.max_rows', 10)
+
+# pandas 3.0 defaults to Arrow-backed strings (future.infer_string=True). The devstack is written
+# for pandas 2.x semantics — df[col].values as numpy object arrays, needed by GEMPACK/harpy HAR I/O
+# and other code that passes .values into numpy-expecting APIs. Restore 2.x behavior devstack-wide.
+# Guarded so it's a harmless no-op on pandas versions that lack this option.
+try:
+    pd.set_option('future.infer_string', False)
+except Exception:
+    pass
 
 import_medium_level = 1 # If this is not true, will import all of the HB library on first import, which can take up to 7 seconds.
 use_strict_importing = 0 
@@ -106,7 +130,17 @@
 from hazelbean.utils import *
 if report_import_times:
     hb.timer('utils')
-
+
+import hazelbean.initialize_definitions
+from hazelbean.initialize_definitions import *
+if report_import_times:
+    hb.timer('initialize_definitions')
+
+import hazelbean.assign_to_object
+from hazelbean.assign_to_object import *
+if report_import_times:
+    hb.timer('assign_to_object')
+
 import hazelbean.arrayframe
 from hazelbean.arrayframe import *
 if report_import_times: