"""The command line. app serve [--host H] [--port P] [--reload] app job pull|extract|embed|score [--payload JSON] app worker [--workers http=8,cpu=1] [--until-idle] app companies seed|add|edit|checked|remove|by-hand|set-board|detach|discover|check|size app searches add|list|remove app agents list|prompt app stages [extract|summarise|judge|fetch_text on|off] app dupes app prep [unmapped|unlinked] [--remap] app db init `app job` queues the work on a queue of its own and drains it here, so a cascade such as `pull` runs every stage it starts before the command returns, and prints one line per stage. `app worker` is the boilerplate's: the same queue with no web app, and the schedules fired from it. `python -m app` is the same thing, for a checkout with no console script. """ from __future__ import annotations import argparse import json import logging import os import sys from pathlib import Path from typing import TYPE_CHECKING, Any from app import config as config_mod if TYPE_CHECKING: from app.models import Subscription from app.storage import Store #: The one plugin this app ships lives outside the `app` package, in #: `plugins/`, so it is discovered by `desk.sources.registry` the same way a #: user's own drop-in plugin would be. `companies` and `job pull` both want #: its plain functions (`is_known_target`, `careers_url`, `fetch_board`) as #: well as the `Source` class, so the repo root goes on `sys.path` once, here, #: rather than each caller guessing where `plugins/` is relative to `cwd`. _ROOT = Path(__file__).resolve().parents[1] if str(_ROOT) not in sys.path: sys.path.insert(0, str(_ROOT)) def _ids(provider: str, table: str, lead: str) -> str: """`lead` and the `id label` pairs of a search reader's `REGIONS`, `MODES` or `REMOTE` for a help text, or nothing when the reader is not loaded or has no such table.""" from app import jobs pairs = ", ".join(f"{key or 'empty'} {label}".replace("%", "%%") for key, label in jobs.sources.provider_table(provider, table).items()) return f"{lead}{pairs}" if pairs else "" def build_parser() -> argparse.ArgumentParser: from app.models import HAND_REASONS from plugins import readers parser = argparse.ArgumentParser( prog="app", description="roledesk: a job-search engine on desk and the pipeline queue.") parser.add_argument("-c", "--config", default=None, help="the YAML file to read (default: $APP_CONFIG, then config.yaml)") parser.add_argument("-v", "--verbose", action="store_true") sub = parser.add_subparsers(dest="command", required=True) serve = sub.add_parser("serve", help="run the web app") serve.add_argument("--host", default="", help="override config server.host") serve.add_argument("--port", type=int, default=0, help="override config server.port") serve.add_argument("--reload", action="store_true", help="restart on a source change") job = sub.add_parser("job", help="queue a job, task or cascade and drain the queue") job.add_argument("kind", help="pull, extract, embed, score, or any registered job") job.add_argument("--payload", default="{}", help="JSON spread over a job handler's keyword arguments") worker = sub.add_parser("worker", help="drain the job table with no web app") worker.add_argument("--workers", default="", help="resource=count,... e.g. cpu=2") worker.add_argument("--until-idle", action="store_true", help="run what is queued, then exit") companies = sub.add_parser("companies", help="the target company list") companies_sub = companies.add_subparsers(dest="action", required=True) seed = companies_sub.add_parser( "seed", help="import seed/companies.json, and the boards it names for rows without one") seed.add_argument("--path", default="seed/companies.json") # One word after `add` is a company for the list checked by hand, two are # a board and its company, which is what `add` meant before that list. add = companies_sub.add_parser( "add", help="add NAME to the list checked by hand, or with BOARD NAME follow a " "company by its board, attaching or replacing an existing one's") add.add_argument("first", metavar="NAME|BOARD", help="the company, or host:slug, e.g. greenhouse:acme, before its name") add.add_argument("second", metavar="NAME", nargs="?") add.add_argument("--careers-url", default="", help="its careers page, http or https") add.add_argument("--note", default="") add.add_argument("--reason", default="added", choices=HAND_REASONS, metavar="REASON", help=f"one of {', '.join(repr(r) for r in HAND_REASONS)}") edit = companies_sub.add_parser("edit", help="change a company's name, link, note or reason") edit.add_argument("key") edit.add_argument("--name") edit.add_argument("--careers-url") edit.add_argument("--note") edit.add_argument("--reason", choices=HAND_REASONS, metavar="REASON", help=f"one of {', '.join(repr(r) for r in HAND_REASONS)}") checked = companies_sub.add_parser( "checked", help="record that its careers page was looked at just now") checked.add_argument("key") remove = companies_sub.add_parser( "remove", help="delete a company with no board on and no open roles") remove.add_argument("key") companies_sub.add_parser( "by-hand", help="the companies checked by hand, never checked first, then the " "longest ago") set_board = companies_sub.add_parser( "set-board", help="attach a board to a company, or replace the one it has") set_board.add_argument("key", help="the company key, e.g. trade-republic") set_board.add_argument("board", help="host:slug, e.g. greenhouse:acme") detach = companies_sub.add_parser( "detach", help="stop reading a company's board and close its roles; the row stays") detach.add_argument("key") discover = companies_sub.add_parser( "discover", help="find a company's board from its website; --add follows it") discover.add_argument("name") discover.add_argument("domain", help="its website, e.g. pleo.io") discover.add_argument("--add", action="store_true", help="add or update the company: its board when one is readable, " "else the host it runs on") companies_sub.add_parser("check", help="report boards that do not answer, now or in the pull") size = companies_sub.add_parser( "size", help="size a company's board again, or --missing every board without a size") size.add_argument("key", nargs="?", default="", help="the company key, e.g. infineon") size.add_argument("--missing", action="store_true", help="every board with no size yet") searches = sub.add_parser("searches", help="saved searches on cross-company job boards") searches_sub = searches.add_subparsers(dest="action", required=True) search_add = searches_sub.add_parser( "add", help="run a search once, then save it for every pull (or update it)") search_add.add_argument("provider", choices=list(readers.PROVIDERS)) search_add.add_argument("name", help='what to call it, e.g. "Copenhagen backend"') search_add.add_argument("--q", required=True, help="the query, in the provider's own syntax") search_add.add_argument("--area", type=int, default=0, help="jobindex geoareaid: 1 Region Hovedstaden, 8 Aarhus") search_add.add_argument("--country", default="", help="adzuna country: gb, de, nl, fr, at, be, es, it, pl, ch; " "freehire countries, comma-separated ISO codes: dk,se, " "empty all") search_add.add_argument("--where", default="", help="adzuna place, e.g. Berlin") search_add.add_argument("--region", default="", help="platsbanken län id: CaRE_1nn_cSU is Skåne, empty all of " "Sweden" + _ids("jobnet", "REGIONS", "; jobnet region: ") + _ids("jobbank", "REGIONS", "; jobbank amt: ")) search_add.add_argument("--remote", default="", help="freehire work mode" + _ids("freehire", "MODES", ": ") + _ids("jobbank", "REMOTE", "; jobbank remote work: ") + ", empty any") search_add.add_argument("--pages", type=int, default=0, help="result pages per pull (default 5, at most 10)") searches_sub.add_parser("list", help="every saved search and its open roles") search_remove = searches_sub.add_parser( "remove", help="stop running a search and close its open roles") search_remove.add_argument("name") agents = sub.add_parser("agents", help="search agents and their prompts") agents_sub = agents.add_subparsers(dest="action", required=True) agents_sub.add_parser("list", help="every agent and its prompts") prompt = agents_sub.add_parser( "prompt", help="add a prompt to an agent, replace its text, or --remove it") prompt.add_argument("agent", help="the agent's name") prompt.add_argument("name", help='the prompt\'s name, e.g. "AI leadership"') prompt.add_argument("text", nargs="?", default="", help="what the agent looks for, in a sentence or a paragraph") prompt.add_argument("--remove", action="store_true", help="remove the prompt instead") stages = sub.add_parser( "stages", help="show whether extract, summarise, judge and fetch_text run, or switch one") stages.add_argument("name", nargs="?", choices=["extract", "summarise", "judge", "fetch_text"]) stages.add_argument("state", nargs="?", choices=["on", "off"]) sub.add_parser("seniority", help="per host, how often the stated level equals the model's") sub.add_parser("dupes", help="recompute which open roles are copies of one posting") prep = sub.add_parser( "prep", help="map the roles' host words and fold the copies, or list what is unmapped") prep.add_argument("action", nargs="?", choices=["unmapped", "unlinked"], help="list the host words no table maps, or the place texts no " "place resolves, instead of running") prep.add_argument("--remap", action="store_true", help="map every role again, closed ones included") db = sub.add_parser("db", help="database housekeeping") db.add_argument("action", choices=["init"]) return parser def main(argv: list[str] | None = None) -> int: args = build_parser().parse_args(argv) logging.basicConfig(level=logging.DEBUG if args.verbose else logging.INFO, format="%(asctime)s %(levelname)s %(name)s: %(message)s") try: if args.command == "serve": return serve(args) if args.command == "job": return run_job(args) if args.command == "worker": return run_worker(args) if args.command == "companies": return run_companies(args) if args.command == "searches": return run_searches(args) if args.command == "agents": return run_agents(args) if args.command == "stages": return run_stages(args) if args.command == "seniority": return run_seniority(args) if args.command == "dupes": return run_dupes(args) if args.command == "prep": return run_prep(args) return init_db(args) except (FileNotFoundError, config_mod.ConfigError) as exc: print(exc, file=sys.stderr) return 1 def serve(args: argparse.Namespace) -> int: import uvicorn config = config_mod.load(args.config) # a config error here, not in uvicorn host = args.host or config.server.host port = args.port or config.server.port level = str(config.log_level).lower() if args.reload: # The reloader builds the app in a fresh process from an import # string, and that process reads the config from $APP_CONFIG. if args.config: os.environ["APP_CONFIG"] = str(args.config) uvicorn.run("app.main:app", host=host, port=port, reload=True, log_level=level) return 0 from app.main import create_app uvicorn.run(create_app(config), host=host, port=port, log_level=level) return 0 def run_job(args: argparse.Namespace) -> int: """Queue `kind` (a cascade such as `pull`, a task from `pipeline.TASKS`, or a job), run everything queued until the queue is idle, and print how it ended: one line per task, from each task's final job. `pull` is a cascade, so it prints a line for every stage it went through. The queue fires no schedules: the command asked for one piece of work.""" from app import jobs from app.storage import Store from pipeline import FINAL, JOBS, TASKS, JobQueue, Jobs cascade = jobs.cascade try: payload: dict[str, Any] = json.loads(args.payload or "{}") except json.JSONDecodeError as exc: print(f"--payload is not JSON: {exc}", file=sys.stderr) return 2 chained = args.kind in cascade.CASCADES if not chained and args.kind not in TASKS and args.kind not in JOBS: known = sorted(set(JOBS) | set(TASKS)) print(f"no job kind {args.kind!r}. Have: {', '.join(known)}", file=sys.stderr) return 2 config = config_mod.load(args.config) store = Store(config.database) queue = JobQueue(Jobs(str(config.database)), config, store, workers=config.workers.model_dump(), bootstrap=jobs.BOOTSTRAP, bootstrap_arg=str(config.path or ""), tz=config.timezone, schedules=False) if chained or args.kind not in JOBS: first = cascade.chain(args.kind) if chained else TASKS[args.kind]() if args.kind == "pull": # The same claim the `pull` job makes, so a pull here and one in # the server's worker cannot both fetch. run = jobs.pull.claim_run(store) if not run: print(json.dumps({"task": "pull", "result": jobs.pull.ALREADY_RUNNING, "running": jobs.pull.running(store)})) return 1 first = cascade.pull_chain(run) task_id = queue.start_task(first) queue.drain() started = queue.jobs.of_task(task_id) if not started: print(json.dumps({"task": args.kind, "result": "nothing to do"})) return 0 whole = queue.jobs.of_cascade(started[0].cascade_id) finals = [j for j in whole if j.kind == FINAL] for final in finals: # A task's own jobs say what the work did; the final job only # tallies how they ended. Up to five, so a pull of twenty board # batches stays one readable line. work = [j.result for j in whole if j.task_id == final.task_id and j.kind != FINAL][:5] print(json.dumps({"id": final.id, "state": final.state, "result": final.result, "work": work, "error": final.error})) return 0 if all(f.state == "done" for f in finals) else 1 queued = queue.enqueue(args.kind, payload) queue.drain() if queued is not None: done = queue.jobs.get(queued.id) if done is not None: print(json.dumps({"id": done.id, "state": done.state, "result": done.result, "error": done.error})) return 0 if done.state == "done" else 1 return 0 def run_worker(args: argparse.Namespace) -> int: """`python -m pipeline.worker` with this app's settings filled in: the bootstrap, the pools from `workers:`, and the timezone, without which a `daily at 06:00` schedule would run at 06:00 on the machine's clock.""" from app.jobs import BOOTSTRAP from pipeline.worker import main as worker_main config = config_mod.load(args.config) argv = ["--db", str(config.database), "--bootstrap", BOOTSTRAP, "--bootstrap-arg", str(args.config or config.path or ""), "--timezone", str(config.timezone)] argv += ["--workers", args.workers or ",".join(f"{name}={n}" for name, n in config.workers.model_dump().items())] if args.until_idle: argv.append("--until-idle") return worker_main(argv) def run_companies(args: argparse.Namespace) -> int: from app.storage import Store config = config_mod.load(args.config) store = Store(config.database) if args.action == "seed": return seed_companies(store, args.path) if args.action == "add": return add_company(store, args) if args.action == "edit": return edit_company(store, args) if args.action == "checked": return company_checked(store, args.key) if args.action == "remove": return remove_company(store, args.key) if args.action == "by-hand": return list_by_hand(store) if args.action == "set-board": return set_board(store, args.key, args.board) if args.action == "detach": return detach_board(store, args.key) if args.action == "discover": return discover_company(store, args.name, args.domain, add=args.add) if args.action == "size": return size_boards(store, args.key, missing=args.missing) return check_companies(store) def seed_companies(store: Store, path: str) -> int: """`app companies seed`: `app.jobs.sources.seed_companies`, its counts as JSON and each entry that did not land whole on stderr.""" from app import jobs result = jobs.sources.seed_companies(store, path) for problem in result.pop("problems"): print(f"companies seed: {problem}", file=sys.stderr) print(json.dumps(result)) return 0 def discover_company(store: Store, name: str, domain: str, *, add: bool) -> int: """`app companies discover "Pleo" pleo.io [--add]`: what the company's site says its board is, as JSON. With `--add`, the company is added or updated: its board attached or replaced when one is readable, else the host and slug it runs on recorded; see `add_discovered`.""" from app import jobs found = jobs.discover.discover(name, domain) out: dict[str, Any] = {"name": name, "domain": domain, "board": found.board, "ats": found.ats, "ats_slug": found.ats_slug, "careers_url": found.careers_url, "evidence": found.evidence} if add: try: out["key"], out["roles"] = jobs.sources.add_discovered(store, name, domain, found) except ValueError as exc: print(f"companies discover: {exc}", file=sys.stderr) return 1 print(json.dumps(out)) return 0 def add_company(store: Store, args: argparse.Namespace) -> int: """`app companies add "Acme" [--careers-url URL] [--note TEXT] [--reason REASON]` puts a company with no board on the list checked by hand; `app companies add greenhouse:acme "Acme"` follows it by its board.""" if args.second is not None: return add_company_by_board(store, args.first, args.second) from app import jobs try: company = jobs.sources.add_by_hand(store, args.first, careers_url=args.careers_url, notes=args.note, hand_reason=args.reason) except ValueError as exc: print(f"companies add: {exc}", file=sys.stderr) return 1 print(json.dumps({"key": company.key, "hand_reason": company.hand_reason})) return 0 def add_company_by_board(store: Store, board: str, name: str) -> int: """`app companies add greenhouse:acme "Acme"`: fetch the board first and refuse if it does not answer or is not a known host. A company that exists gets the board attached or replaced. Port of seedgen's `-add`.""" from app import jobs from app.models import company_key try: sub, found = jobs.sources.add_board(store, board, name) except ValueError as exc: print(f"companies add: {exc}", file=sys.stderr) return 1 print(json.dumps({"key": company_key(sub.company), "board": sub.target, "roles": found})) return 0 def edit_company(store: Store, args: argparse.Namespace) -> int: """`app companies edit acme --careers-url https://acme.example/jobs`: only the options given change; see `jobs.sources.edit_company`.""" from app import jobs given = {"name": args.name, "careers_url": args.careers_url, "notes": args.note, "hand_reason": args.reason} try: company = jobs.sources.edit_company( store, args.key, **{f: v for f, v in given.items() if v is not None}) except KeyError: print(f"companies edit: no company {args.key!r}", file=sys.stderr) return 1 except ValueError as exc: print(f"companies edit: {exc}", file=sys.stderr) return 1 print(json.dumps({"key": company.key, "name": company.name, "careers_url": company.careers_url, "notes": company.notes, "hand_reason": company.hand_reason})) return 0 def company_checked(store: Store, key: str) -> int: """`app companies checked acme`: its careers page was looked at just now, which the Watchlist keeps as `careers_opened_at`, the same stamp opening the page from there sets.""" if store.company(key) is None: print(f"companies checked: no company {key!r}", file=sys.stderr) return 1 company = store.update_company(key, careers_opened_at=store.now()) print(json.dumps({"key": key, "careers_opened_at": company.careers_opened_at})) return 0 def remove_company(store: Store, key: str) -> int: """`app companies remove acme`: refused for a company still read, see `jobs.sources.remove_company`.""" from app import jobs try: jobs.sources.remove_company(store, key) except KeyError: print(f"companies remove: no company {key!r}", file=sys.stderr) return 1 except ValueError as exc: print(f"companies remove: {exc}", file=sys.stderr) return 1 print(json.dumps({"key": key, "removed": True})) return 0 def list_by_hand(store: Store) -> int: """`app companies by-hand`: the companies with no board shown that are not known from a search only, one line each, never opened first and then the longest ago; `watch` is the Watchlist's rating, yes, no or a dash for unrated. See `jobs.sources.by_hand`.""" from app import jobs rows = jobs.sources.by_hand(store) for c in rows: watch = "-" if c.watch is None else "yes" if c.watch else "no" opened = c.careers_opened_at[:10] or "never" print(f"{c.name:30} {c.hand_reason or '-':8} {watch:4} {opened:10} {c.careers_url}") print(f"{len(rows)} by hand") return 0 def set_board(store: Store, key: str, board: str) -> int: """`app companies set-board trade-republic greenhouse:traderepublic`: attach the board to an existing company or replace the one it has.""" from app import jobs try: sub, found, closed = jobs.sources.set_board(store, key, board) except ValueError as exc: print(f"companies set-board: {exc}", file=sys.stderr) return 1 print(json.dumps({"key": key, "board": sub.target, "roles": found, "closed": closed})) return 0 def detach_board(store: Store, key: str) -> int: """`app companies detach acme`: stop reading its board, close its roles.""" from app import jobs try: closed = jobs.sources.detach_board(store, key) except KeyError: print(f"companies detach: no company {key!r}", file=sys.stderr) return 1 except ValueError as exc: print(f"companies detach: {exc}", file=sys.stderr) return 1 print(json.dumps({"key": key, "closed": closed})) return 0 def size_boards(store: Store, key: str, *, missing: bool) -> int: """`app companies size infineon`: size its board again and keep it, as the Companies row's Recheck does, and print the size and what a pull asks the board for. `--missing` sizes every board shown that has no size yet, a line each, and exits 1 if any did not answer. The next pull reads a board so; nothing is queued from here.""" from app import jobs if bool(key) == missing: print("companies size: name a company key, or --missing", file=sys.stderr) return 2 if missing: subs = [s for s in store.subscriptions(source="ats") if s.show] known = jobs.reach.sizes(store, subs) subs = [s for s in subs if s.target not in known] else: sub = store.company_subscription(key) if sub is None or not sub.show: print(f"companies size: {key!r} has no board", file=sys.stderr) return 1 subs = [sub] sized: list[tuple[Subscription, int]] = [] for sub in subs: try: sized.append((sub, jobs.sources.size_board(store, sub.target))) except ValueError as exc: print(f"companies size: {exc}", file=sys.stderr) plan = jobs.reach.plan(store, [sub for sub, _size in sized]) for sub, size in sized: print(json.dumps({"board": sub.target, "size": size, "reach": jobs.reach.line(plan, sub)})) dead = len(subs) - len(sized) if missing: print(f"sized {len(sized)} of {len(subs)} boards without a size") return 1 if dead else 0 def check_companies(store: Store) -> int: """`app companies check`: every `ats` board, fetched once. Prints a line per board that did not answer now, answered incompletely, or has been failing in the pull, with what the pull recorded beside the live result, and exits 1 if any did not answer now.""" from app import jobs from plugins.ats import failure_reason, fetch_board subs = [s for s in store.subscriptions(source="ats") if s.show] failing = jobs.pull.failing_boards(store) dead = 0 for sub in subs: try: rows, complete = fetch_board(sub.target) live = f"{len(rows)} roles" + ("" if complete else ", incomplete") except Exception as exc: complete, live = False, f"not answering: {failure_reason(exc)}" dead += 1 since, reason = failing.get(sub.target, ("", "")) if complete and not since: continue stored = f"pull: failing since {since[:10]}: {reason}" if since else "pull: answering" print(f"{sub.target} ({sub.label}): now {live}; {stored}") print(f"checked {len(subs)} boards, {dead} dead, {len(failing)} failing in the pull") return 1 if dead else 0 def run_searches(args: argparse.Namespace) -> int: from app import jobs from app.storage import Store config = config_mod.load(args.config) store = Store(config.database) if args.action == "add": return add_search(config, store, args) if args.action == "remove": sub = store.subscription_for("search", args.name) if sub is None: print(f"searches remove: no search named {args.name!r}", file=sys.stderr) return 1 _, closed = jobs.sources.remove_search(store, sub.id) print(json.dumps({"removed": sub.label, "roles_closed": closed})) return 0 countries = store.open_countries_by_subscription() for sub in store.subscriptions(source="search"): removed = "" if sub.subscribed and sub.fetch else " (off)" open_roles = sum(countries.get(sub.id, {}).values()) print(f"{sub.label:30} {sub.target:40} {open_roles:5} open " f"{json.dumps(sub.params, ensure_ascii=False)}{removed}") return 0 def add_search(config: config_mod.Config, store: Store, args: argparse.Namespace) -> int: """`app searches add jobindex "Copenhagen backend" --q '"backend developer"' --area 1`: run the search once and save it only if it finds something. A search of the same name is updated in place.""" from app import jobs try: params = jobs.sources.search_params(args.provider, args.q, area=args.area, country=args.country, where=args.where, region=args.region, remote=args.remote, pages=args.pages) sub, found, complete = jobs.sources.save_search(config, store, args.provider, args.name, params) except ValueError as exc: print(f"searches add: {exc}", file=sys.stderr) return 1 print(json.dumps({"search": sub.label, "target": sub.target, "roles": found, "complete": complete}, ensure_ascii=False)) return 0 def run_agents(args: argparse.Namespace) -> int: """`app agents list` prints every agent and its prompts. `app agents prompt "Wolf, next role" "AI leadership" "Head of AI ..."` adds that prompt at the end, or replaces the text of the one of that name in its place; `--remove` removes it. The agent's other fields stay as they are. Scores follow at the next `score`.""" from app.models import Prompt from app.storage import Store config = config_mod.load(args.config) store = Store(config.database) if args.action == "list": for each in store.agents(): print(each.name + ("" if each.enabled else " (off)")) for p in each.prompts: text = " ".join(p.text.split()) print(f" {p.name}: {text[:100] + '…' if len(text) > 100 else text}") return 0 agent = next((a for a in store.agents() if a.name == args.agent), None) if agent is None: print(f"agents prompt: no agent named {args.agent!r}", file=sys.stderr) return 1 named = args.name.strip().casefold() prompts = [p for p in agent.prompts if not (args.remove and p.name.casefold() == named)] if args.remove and len(prompts) == len(agent.prompts): print(f"agents prompt: {agent.name!r} has no prompt named {args.name!r}", file=sys.stderr) return 1 if not args.remove: try: new = Prompt(name=args.name, text=args.text) except ValueError as exc: print(f"agents prompt: {exc}", file=sys.stderr) return 1 at = next((n for n, p in enumerate(prompts) if p.name.casefold() == named), None) if at is None: prompts.append(new) else: prompts[at] = new saved = store.save_agent(agent.name, gates=agent.gates, weights=agent.weights, enabled=agent.enabled, agent_id=agent.id, limits=agent.limits, prompts=prompts) print(json.dumps({"agent": saved.name, "prompts": [p.name for p in saved.prompts]}, ensure_ascii=False)) return 0 def run_stages(args: argparse.Namespace) -> int: """`app stages` prints what `GET /api/stages` returns; `app stages extract on` stores the switch through the function `PUT /api/stages/extract` calls, and prints the result.""" from app import jobs from app.storage import Store if bool(args.name) != bool(args.state): print("stages: give a stage and on or off, or neither", file=sys.stderr) return 1 config = config_mod.load(args.config) store = Store(config.database) if args.name: jobs.stages.set_stage(store, args.name, args.state == "on") found = [jobs.stages.describe_one(config, store, args.name)] else: found = jobs.stages.describe(config, store) for stage in found: where = "set by the switch" if stage.source == "stored" else "config.yaml's default" print(f"{stage.name:10} {'on ' if stage.on else 'off'} {where}; " f"{stage.model}; {stage.answer}") return 0 def run_seniority(args: argparse.Namespace) -> int: """`app seniority`: per host, the roles that have both a stated and a model level and how many of them agree, `Store.seniority_agreement`, as a table. For deciding which host's stated level to trust.""" from app.storage import Store config = config_mod.load(args.config) found = Store(config.database).seniority_agreement() if not found: print("no role has a stated level yet") return 0 print(f"{'host':20} {'compared':>8} {'agreed':>7} {'rate':>6} {'unread':>7}") for h in found: rate = "-" if h["rate"] is None else f"{h['rate']:.0%}" print(f"{h['host'] or '(none)':20} {h['compared']:8} {h['agreed']:7} {rate:>6} " f"{h['stated_only']:7}") return 0 def run_dupes(args: argparse.Namespace) -> int: """`app dupes`: what the prep stage ends with, `Store.rebuild_dupes`, now. Prints the groups (kept roles with a copy), the copies the feed hides and the rows whose `dup_of` changed, as JSON.""" from app.storage import Store config = config_mod.load(args.config) counts = Store(config.database).rebuild_dupes() print(json.dumps(counts._asdict())) return 0 def run_prep(args: argparse.Namespace) -> int: """`app prep`: the prep stage now (`app.prep.run`), printing what it did as JSON; `--remap` maps every role again. `app prep unmapped` lists the host words no table maps, per field and host, most roles first; `app prep unlinked` the place texts of open roles no place resolves, with the countries and companies of their roles.""" from app import prep from app.storage import Store config = config_mod.load(args.config) store = Store(config.database) if args.action == "unmapped": found = store.unmapped() if not found: print("every host word is mapped") for w in found: print(f"{w['field']:11} {w['host'] or '(none)':16} {w['roles']:6} {w['word']}") return 0 if args.action == "unlinked": linked = store.unlinked_places() if not linked: print("every place text is linked") for p in linked: where = " ".join(f"{c or '-'}:{n}" for c, n in p["countries"][:3]) print(f"{p['roles']:6} {p['text'][:40]:40} {where:20} {', '.join(p['companies'])}") return 0 print(json.dumps(prep.run(store, remap=args.remap)._asdict())) return 0 def init_db(args: argparse.Namespace) -> int: """Create the file, or bring it to the current schema.""" from app.storage import LATEST, Store config = config_mod.load(args.config) store = Store(config.database) applied = store.migrations_applied print(f"{store.path}: schema v{LATEST}" + (f", applied {applied}" if applied else "")) return 0 if __name__ == "__main__": raise SystemExit(main())