Skip to main content

Deterministic Critical Chain Project Management (CCPM) scheduler: validate a project network, build a resource-leveled buffered schedule, verify it, and plot a buffer-aware Gantt chart. Library + CLI, designed for both humans and AI agents.

Project description

ccpm-scheduler

Deterministic Critical Chain Project Management (CCPM) scheduling as a Python library and CLI — for humans, for AI agents, and for embedding in other tools.

Give it a project network (tasks, dependencies, duration estimates, resource assignments) plus resource availability, and it produces a proper Critical Chain schedule: resource-leveled, scheduled as late as possible, protected by a project buffer and feeding buffers, with a buffer-aware Gantt chart.

Same input always yields byte-identical output — the scheduler is fully deterministic, so it is scriptable, diffable, and testable.

Status

Phases 1–2 complete: extracted behavior-preserving from the ccpm-scheduler Claude skill (guarded by byte-identical golden tests), with a typed model, library API, and the ccpm-scheduler CLI. Coming next (see PLAN.md): the Claude skill drives this CLI (Phase 3), and the our-planner GUI imports the library (Phase 4).

CLI

Built for humans and AI agents: exit codes are a contract (0 = ok, 1 = problems found with the report still emitted, 2 = usage error), --json prints a machine-readable document, there are no interactive prompts, and the same input always produces byte-identical output.

ccpm-scheduler validate tasks.csv resources.csv calendar.csv
ccpm-scheduler build tasks.csv resources.csv --calendar calendar.csv \
    --out-dir plan --title "Website relaunch"
ccpm-scheduler check plan/schedule.csv tasks.csv resources.csv calendar.csv
ccpm-scheduler plot plan/schedule.csv plan/gantt.png --resources resources.csv
ccpm-scheduler graph plan/schedule.csv plan/project-network.html \
    --title "Website relaunch"   # interactive dependency graph
ccpm-scheduler schema network     # JSON Schema of the input format

graph writes a standalone interactive network view of the schedule (vis-network from a CDN, data embedded — no build step or server): open the HTML in any browser to zoom, pan, drag nodes, toggle hierarchical/free layout, and click a node to inspect its schedule, resources, and links. The Gantt shows when; the graph shows why — colors match the Gantt (critical chain firebrick, feeding chains colored, buffers gold/khaki with dashed attachment edges).

The network input is either CSV files or a single JSON document (a path, or - for stdin) in the exchange format:

echo '{"tasks": [...], "resources": [...]}' | ccpm-scheduler build - --json

build validates first: on a broken network you get the same coded issue report as validate (exit 1) and no files. --json reports carry stable machine-readable issue codes (E_CYCLE, E_NO_RESOURCE, E_FRACTIONAL_ALLOCATION, …) plus the offending task/resource ids; ccpm-scheduler schema report describes the shape.

Library API

from ccpm_scheduler import (load_network, validate_network,
                            build_schedule, check_schedule, plot_schedule)

network = load_network("tasks.csv", "resources.csv", "calendar.csv")
report = validate_network(network)   # ValidationReport with coded Issues
if report.ok:
    result = build_schedule(network, title="My project")
    assert check_schedule(result.schedule, network).ok
    plot_schedule(result.schedule, "gantt.png",
                  resources=network.resources, calendar=network.calendar)
    print(result.stats.status_line("My project"))

Validation issues carry machine-readable codes (E_CYCLE, E_NO_RESOURCE, E_FRACTIONAL_ALLOCATION, …) plus the offending task/resource ids, so an embedding tool can annotate its own UI. Networks also load from a JSON exchange format (network_from_json / network_to_json) that accepts structured predecessor links ({"id": 2, "type": "SS", "lag": 2}), numeric ids, and per-resource allocation maps — the shape GUI tools naturally emit. Fractional allocations and capacities are rejected with precise errors in v1 (whole resources only).

(python -m ccpm_scheduler is equivalent to the ccpm-scheduler command; the individual stages also remain runnable as python -m ccpm_scheduler.validate etc. with their original argument conventions.)

Input contract

tasks.csvid, name, realistic_duration, optimal_duration (optional), predecessor_ids, resource_ids, url (optional)

  • realistic_duration: estimate with safety included; optimal_duration: padding-free estimate. If optimal_duration is missing, the classic 50% cut is applied. (Legacy names duration_safe/duration_aggressive are accepted.)
  • predecessor_ids: semicolon-separated links — bare id = Finish-to-Start, typed links with lag supported: A:SS+2, A:FF, A:SF. Buffer rows in the output use the CCPM-specific :PB/:FB link types.
  • resource_ids: semicolon-separated; every task needs at least one resource.

resources.csvid, name, capacity, url (optional); capacity defaults to 1.

calendar.csv (optional) — resource_id, from, to, capacity: overrides a resource's capacity on the half-open day range [from, to); capacity = 0 means unavailable. Tasks execute contiguously — they never pause across an outage.

Durations are working days; the schedule uses integer day offsets from day 0.

Output

  • schedule.csv — tasks and buffers with start/finish, chain membership (critical, feeding-n), and link notation (buffers attached via :PB/:FB)
  • summary.md — critical chain, project duration, buffer sizes, promised completion date (= end of the project buffer)
  • gantt.png — critical chain, feeding chains, buffers, dependency arrows, and a resource-utilization panel on the same time axis
  • project-network.html (via graph) — standalone interactive dependency graph of the same schedule, for exploring the network structure the Gantt can't show

Development

uv sync
uv run pytest

The golden tests in tests/ assert byte-identical schedule.csv/summary.md for five reference projects. Regenerate goldens only for deliberate, reviewed behavior changes.

License

Dual-licensed under either of MIT or Apache License 2.0, at your option.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

ccpm_scheduler-0.5.0.tar.gz (128.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

ccpm_scheduler-0.5.0-py3-none-any.whl (48.3 kB view details)

Uploaded Python 3

File details

Details for the file ccpm_scheduler-0.5.0.tar.gz.

File metadata

  • Download URL: ccpm_scheduler-0.5.0.tar.gz
  • Upload date:
  • Size: 128.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: uv/0.11.28 {"installer":{"name":"uv","version":"0.11.28","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for ccpm_scheduler-0.5.0.tar.gz
Algorithm Hash digest
SHA256 44c232c94a347c1884fe51aff19600314ae564366de6919946240dc1ff326f93
MD5 07ba0f4e3b77ba6f3d3d714d3fc1287b
BLAKE2b-256 15b719a296773c8caa0c3f108a0b2076d8f2cb35c0370ebe292a04016e609ee8

See more details on using hashes here.

File details

Details for the file ccpm_scheduler-0.5.0-py3-none-any.whl.

File metadata

  • Download URL: ccpm_scheduler-0.5.0-py3-none-any.whl
  • Upload date:
  • Size: 48.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: uv/0.11.28 {"installer":{"name":"uv","version":"0.11.28","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for ccpm_scheduler-0.5.0-py3-none-any.whl
Algorithm Hash digest
SHA256 8137d95ea824fd2015df8cc46f33dcb7305067abd2a80568f93dd6ddc6ac98c6
MD5 80027c59d4df3ca58ef67af5a3f01440
BLAKE2b-256 cf6c5dc83c9b6bd0f63097608f675a0562b9aaa08ceda913e4b9063ff48ee47f

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page