stage-cli 1.0.0__py3-none-any.whl

stage/data/lexicon/terms.yaml

@@ -0,0 +1,76 @@
+summer:
+  - summer
+  - été
+  - ete
+  - estival
+  - estivale
+  - summer term
+  - session d'été
+  - stage d'été
+
+fall:
+  - fall
+  - autumn
+  - automne
+  - fall term
+  - session d'automne
+  - stage d'automne
+
+winter:
+  - winter
+  - hiver
+  - hivernale
+  - winter term
+  - session d'hiver
+  - stage d'hiver
+
+spring:
+  - spring
+  - printemps
+  - spring term
+  - session du printemps
+  - stage du printemps
+
+fillers:
+  - of
+  - de
+  - du
+  - des
+  - d
+  - the
+  - la
+  - le
+  - term
+  - session
+  - start
+  - starting
+  - intake
+  - cohort
+  - programme
+  - program
+
+blocked_bigrams:
+  - spring boot
+  - spring framework
+  - spring frameworks
+  - spring security
+  - spring mvc
+  - spring data
+  - spring cloud
+  - spring batch
+  - spring actuator
+  - spring health
+  - spring environments
+  - spring integration
+  - spring webflux
+  - summer camp
+  - winter garden
+  - fall protection
+  - fall risk
+  - fall arrest
+  - spring loaded
+  - spring valley
+  - winter park
+  - summer street
+  - spring street
+  - spring hill

stage/data/lexicon/workday_facets.yaml

@@ -0,0 +1,27 @@
+facet_parameters:
+  - timeType
+  - workerSubType
+  - jobFamily
+  - employmentType
+
+intern_descriptors:
+  - intern
+  - interns
+  - internship
+  - internships
+  - intern/co-op
+  - co-op
+  - coop
+  - co-op/intern
+  - co-op student
+  - stage
+  - stagiaire
+  - stage etudiant
+  - stagiaire etudiant
+  - student
+  - student worker
+  - etudiant
+  - alternance
+  - alternant
+  - summer intern
+  - summer student

stage/data/seed_companies.yaml

@@ -0,0 +1,198 @@
+- { name: Behaviour Interactive, section: montreal, priority: high }
+- { name: Bombardier, section: montreal, priority: high }
+- { name: CAE, section: montreal, priority: high }
+- { name: Coveo, section: montreal, priority: high }
+- { name: Desjardins, section: montreal, priority: high }
+- { name: Eidos-Montréal, section: montreal, priority: high }
+- { name: Gameloft, section: montreal, priority: high }
+- { name: Genetec, section: montreal, priority: high }
+- { name: Hydro-Québec, section: montreal, priority: high }
+- { name: Intact, section: montreal, priority: high }
+- { name: Laurentian Bank, section: montreal, priority: high }
+- { name: Lightspeed, section: montreal, priority: high }
+- { name: Matrox, section: montreal, priority: high }
+- { name: Mila, section: montreal, priority: high }
+- { name: National Bank of Canada, section: montreal, priority: high }
+- { name: Nuvei, section: montreal, priority: high, note: not listed, consider }
+- { name: Pratt & Whitney Canada, section: montreal, priority: high }
+- { name: Squarepoint Capital, section: montreal, priority: high }
+- { name: Ubisoft, section: montreal, priority: high }
+- { name: Valsoft, section: montreal, priority: high }
+- { name: Videotron, section: montreal, priority: high }
+- { name: Warner Bros Games, section: montreal, priority: high }
+
+- { name: BlackBerry, section: canada, priority: high }
+- { name: BlackBerry QNX, section: canada, priority: high }
+- { name: BMO, section: canada, priority: high }
+- { name: Bell, section: canada, priority: high }
+- { name: CGI, section: canada, priority: high }
+- { name: CIBC, section: canada, priority: high }
+- { name: Ciena, section: canada, priority: high }
+- { name: Interac, section: canada, priority: high }
+- { name: Mackenzie Investments, section: canada, priority: high }
+- { name: Manulife, section: canada, priority: high }
+- { name: MDA Space, section: canada, priority: high }
+- { name: OpenText, section: canada, priority: high, note: not listed, consider }
+- { name: RBC, section: canada, priority: high }
+- { name: Rogers, section: canada, priority: high }
+- { name: Scotiabank, section: canada, priority: high }
+- { name: Shopify, section: canada, priority: high }
+- { name: Sun Life, section: canada, priority: high }
+- { name: TD Bank, section: canada, priority: high }
+- { name: Telus, section: canada, priority: high }
+- { name: Wealthsimple, section: canada, priority: high }
+
+- { name: Akuna Capital, section: quant, priority: normal }
+- { name: Citadel, section: quant, priority: normal }
+- { name: Citadel Securities, section: quant, priority: normal }
+- { name: D.E. Shaw, section: quant, priority: normal }
+- { name: DRW, section: quant, priority: normal }
+- { name: Hudson River Trading, section: quant, priority: normal }
+- { name: Jane Street, section: quant, priority: normal }
+- { name: Jump Trading, section: quant, priority: normal }
+- { name: Maven Securities, section: quant, priority: normal }
+- { name: Optiver, section: quant, priority: normal }
+- { name: Renaissance Technologies, section: quant, priority: low, note: does not recruit publicly }
+- { name: Tower Research Capital, section: quant, priority: normal }
+- { name: Two Sigma, section: quant, priority: normal }
+
+- { name: Adobe, section: bigtech, priority: normal }
+- { name: Amazon, section: bigtech, priority: normal }
+- { name: AMD, section: bigtech, priority: normal }
+- { name: Apple, section: bigtech, priority: normal }
+- { name: Broadcom, section: bigtech, priority: normal }
+- { name: ByteDance, section: bigtech, priority: normal }
+- { name: Cisco, section: bigtech, priority: normal }
+- { name: Cisco Meraki, section: bigtech, priority: normal }
+- { name: Cisco ThousandEyes, section: bigtech, priority: normal }
+- { name: Dell, section: bigtech, priority: normal }
+- { name: Ericsson, section: bigtech, priority: normal }
+- { name: Google, section: bigtech, priority: normal }
+- { name: Hewlett Packard Enterprise, section: bigtech, priority: normal }
+- { name: Hitachi, section: bigtech, priority: low, note: low yield for a Montreal student }
+- { name: Huawei, section: bigtech, priority: normal }
+- { name: IBM, section: bigtech, priority: normal }
+- { name: Intel, section: bigtech, priority: normal }
+- { name: LinkedIn, section: bigtech, priority: normal }
+- { name: Meta, section: bigtech, priority: normal }
+- { name: Microsoft, section: bigtech, priority: normal }
+- { name: Motorola, section: bigtech, priority: low, note: low yield for a Montreal student }
+- { name: Netflix, section: bigtech, priority: normal }
+- { name: Nokia, section: bigtech, priority: normal }
+- { name: NVIDIA, section: bigtech, priority: normal }
+- { name: Oracle, section: bigtech, priority: normal }
+- { name: Qualcomm, section: bigtech, priority: normal, note: not listed }
+- { name: Salesforce, section: bigtech, priority: normal }
+- { name: Samsung, section: bigtech, priority: normal }
+- { name: SAP, section: bigtech, priority: normal }
+- { name: Slack, section: bigtech, priority: normal }
+- { name: Sony Interactive Entertainment, section: bigtech, priority: normal }
+- { name: TikTok, section: bigtech, priority: normal }
+- { name: Tesla, section: bigtech, priority: normal }
+- { name: Uber, section: bigtech, priority: normal }
+- { name: VMware, section: bigtech, priority: normal }
+- { name: X (Twitter), section: bigtech, priority: normal }
+- { name: Zoom, section: bigtech, priority: normal }
+
+- { name: American Express, section: finance, priority: normal }
+- { name: Bank of America, section: finance, priority: normal }
+- { name: BlackRock, section: finance, priority: normal }
+- { name: Bloomberg, section: finance, priority: normal }
+- { name: BNP Paribas, section: finance, priority: normal }
+- { name: Capital One, section: finance, priority: normal }
+- { name: Citigroup, section: finance, priority: normal }
+- { name: Fidelity Investments, section: finance, priority: normal }
+- { name: Goldman Sachs, section: finance, priority: normal }
+- { name: JPMorgan Chase, section: finance, priority: normal }
+- { name: Mastercard, section: finance, priority: normal }
+- { name: Morgan Stanley, section: finance, priority: normal }
+- { name: PayPal, section: finance, priority: normal }
+- { name: Visa, section: finance, priority: normal }
+- { name: Wells Fargo, section: finance, priority: normal }
+
+- { name: Accenture, section: consulting, priority: normal, note: not listed, consider }
+- { name: Deloitte, section: consulting, priority: normal }
+- { name: KPMG, section: consulting, priority: normal }
+- { name: PwC, section: consulting, priority: normal }
+- { name: Red Hat, section: consulting, priority: normal }
+- { name: Splunk, section: consulting, priority: normal }
+
+- { name: Analog Devices, section: industrial, priority: normal }
+- { name: Anduril Industries, section: industrial, priority: normal }
+- { name: BAE Systems, section: industrial, priority: normal }
+- { name: Ford, section: industrial, priority: low, note: low yield for a Montreal student }
+- { name: General Motors, section: industrial, priority: low, note: low yield for a Montreal student }
+- { name: Lockheed Martin, section: industrial, priority: normal }
+- { name: SpaceX, section: industrial, priority: normal }
+- { name: Waymo, section: industrial, priority: normal }
+
+- { name: 1Password, section: security, priority: normal }
+- { name: Akamai, section: security, priority: normal }
+- { name: Arctic Wolf, section: security, priority: normal }
+- { name: Cloudflare, section: security, priority: normal }
+- { name: CrowdStrike, section: security, priority: normal }
+- { name: Fortinet, section: security, priority: normal }
+- { name: Okta, section: security, priority: normal, note: not listed }
+- { name: Palo Alto Networks, section: security, priority: normal }
+- { name: Wiz, section: security, priority: normal }
+
+- { name: Affirm, section: startups, priority: normal }
+- { name: Airbnb, section: startups, priority: normal }
+- { name: Atlassian, section: startups, priority: normal }
+- { name: Autodesk, section: startups, priority: normal }
+- { name: Bumble, section: startups, priority: normal }
+- { name: Canva, section: startups, priority: normal }
+- { name: CircleCI, section: startups, priority: normal }
+- { name: Coinbase, section: startups, priority: normal }
+- { name: Databricks, section: startups, priority: normal }
+- { name: Datadog, section: startups, priority: normal }
+- { name: DigitalOcean, section: startups, priority: normal }
+- { name: Discord, section: startups, priority: normal }
+- { name: Docker, section: startups, priority: normal }
+- { name: DoorDash, section: startups, priority: normal }
+- { name: Dropbox, section: startups, priority: normal }
+- { name: Duolingo, section: startups, priority: normal }
+- { name: eBay, section: startups, priority: normal }
+- { name: Expedia, section: startups, priority: normal }
+- { name: Figma, section: startups, priority: normal }
+- { name: GitHub, section: startups, priority: normal }
+- { name: Grammarly, section: startups, priority: normal }
+- { name: HashiCorp, section: startups, priority: normal }
+- { name: Hugging Face, section: startups, priority: normal }
+- { name: Instacart, section: startups, priority: normal }
+- { name: Lyft, section: startups, priority: normal }
+- { name: MongoDB, section: startups, priority: normal }
+- { name: Notion, section: startups, priority: normal }
+- { name: Palantir, section: startups, priority: normal }
+- { name: Pinterest, section: startups, priority: normal }
+- { name: Postman, section: startups, priority: normal }
+- { name: Quora, section: startups, priority: normal }
+- { name: Reddit, section: startups, priority: normal }
+- { name: Roblox, section: startups, priority: normal }
+- { name: Scale AI, section: startups, priority: normal }
+- { name: Snowflake, section: startups, priority: normal }
+- { name: Spotify, section: startups, priority: normal }
+- { name: Stripe, section: startups, priority: normal }
+- { name: Supabase, section: startups, priority: normal }
+- { name: Twilio, section: startups, priority: normal }
+- { name: Unity, section: startups, priority: normal }
+- { name: Wayfair, section: startups, priority: normal }
+- { name: Yelp, section: startups, priority: normal }
+- { name: Zillow, section: startups, priority: normal }
+
+- { name: Activision Blizzard, section: gaming, priority: normal }
+- { name: Bandai Namco, section: gaming, priority: low, note: negligible North American CS pipeline }
+- { name: Blackbird Interactive, section: gaming, priority: normal }
+- { name: Bungie, section: gaming, priority: normal }
+- { name: Capcom, section: gaming, priority: low, note: negligible North American CS pipeline }
+- { name: CD Projekt Red, section: gaming, priority: normal }
+- { name: EA, section: gaming, priority: normal }
+- { name: Epic Games, section: gaming, priority: normal }
+- { name: FromSoftware, section: gaming, priority: low, note: negligible North American CS pipeline }
+- { name: Halo Studios, section: gaming, priority: normal }
+- { name: Microsoft Xbox Game Studios, section: gaming, priority: normal }
+- { name: Nintendo, section: gaming, priority: low, note: negligible North American CS pipeline }
+- { name: Riot Games, section: gaming, priority: normal }
+- { name: Square Enix, section: gaming, priority: low, note: negligible North American CS pipeline }
+- { name: Supercell, section: gaming, priority: low, note: negligible North American CS pipeline }
+- { name: Take-Two Interactive, section: gaming, priority: normal }

stage/dedup/__init__.py

@@ -0,0 +1,19 @@
+from stage.dedup.identity import (
+    SOURCE_PRIORITY,
+    MatchKind,
+    MatchResult,
+    title_canonical,
+    would_merge,
+)
+from stage.dedup.resolve import DuplicateLink, rank, resolve_duplicates
+
+__all__ = [
+    "SOURCE_PRIORITY",
+    "DuplicateLink",
+    "MatchKind",
+    "MatchResult",
+    "rank",
+    "resolve_duplicates",
+    "title_canonical",
+    "would_merge",
+]

stage/dedup/identity.py

@@ -0,0 +1,113 @@
+from dataclasses import dataclass
+from enum import StrEnum
+
+from stage.domain import (
+    SOURCE_PRIORITY,
+    Job,
+    Language,
+    location_agrees,
+    term_agrees,
+)
+from stage.lexicon import (
+    division_qualifiers,
+    fold,
+    generic_company_tokens,
+    name_root_tokens,
+    role_lexicon,
+)
+
+__all__ = [
+    "SOURCE_PRIORITY",
+    "MatchKind",
+    "MatchResult",
+    "title_canonical",
+    "would_merge",
+]
+
+
+class MatchKind(StrEnum):
+    NONE = "none"
+    URL = "url"
+    SAME_LANGUAGE = "same-language"
+    CROSS_LANGUAGE = "cross-language"
+
+
+@dataclass(frozen=True, slots=True)
+class MatchResult:
+    kind: MatchKind = MatchKind.NONE
+    evidence: str = ""
+
+    def __bool__(self) -> bool:
+        return self.kind is not MatchKind.NONE
+
+
+def title_canonical(title: str) -> str:
+    folded = fold(title)
+    if not folded:
+        return ""
+    padded = f" {folded} "
+    hits: list[tuple[int, str]] = []
+    for category, phrases in role_lexicon().items():
+        matched = [phrase for phrase in phrases if f" {phrase} " in padded]
+        if matched:
+            hits.append((max(len(phrase) for phrase in matched), category))
+    if not hits:
+        return ""
+    best = max(length for length, _ in hits)
+    winners = sorted(category for length, category in hits if length == best)
+    return winners[0] if len(winners) == 1 else ""
+
+
+def _title_tokens(job: Job) -> frozenset[str]:
+    return frozenset(fold(job.title_raw).split())
+
+
+def _qualified_prefix(shorter: tuple[str, ...], longer: tuple[str, ...]) -> bool:
+    if not shorter or longer[: len(shorter)] != shorter:
+        return False
+    if not set(longer[len(shorter) :]) <= division_qualifiers():
+        return False
+    return bool(set(shorter) - generic_company_tokens())
+
+
+def _company_matches(left: Job, right: Job) -> bool:
+    if fold(left.company) == fold(right.company):
+        return True
+    first, second = name_root_tokens(left.company), name_root_tokens(right.company)
+    if first == second:
+        return True
+    shorter, longer = sorted((first, second), key=len)
+    return _qualified_prefix(shorter, longer)
+
+
+def would_merge(left: Job, right: Job) -> MatchResult:
+    if left.board_key == right.board_key:
+        return MatchResult()
+
+    if left.apply_url_canonical and left.apply_url_canonical == right.apply_url_canonical:
+        return MatchResult(MatchKind.URL, left.apply_url_canonical)
+
+    if not _company_matches(left, right):
+        return MatchResult()
+
+    if not location_agrees(left.location, right.location):
+        return MatchResult()
+
+    left_tokens, right_tokens = _title_tokens(left), _title_tokens(right)
+    if left_tokens and left_tokens == right_tokens:
+        return MatchResult(MatchKind.SAME_LANGUAGE, " ".join(sorted(left_tokens)))
+
+    return _cross_language(left, right)
+
+
+def _cross_language(left: Job, right: Job) -> MatchResult:
+    pair = {left.language, right.language}
+    if Language.EN not in pair or Language.FR not in pair:
+        return MatchResult()
+    if not term_agrees(left.term, right.term):
+        return MatchResult()
+
+    canonical = title_canonical(left.title_raw)
+    if not canonical or canonical != title_canonical(right.title_raw):
+        return MatchResult()
+    return MatchResult(MatchKind.CROSS_LANGUAGE, canonical)

stage/dedup/resolve.py

@@ -0,0 +1,97 @@
+from collections.abc import Sequence
+from dataclasses import dataclass
+
+from stage.dedup.identity import MatchKind, would_merge
+from stage.domain import Job, source_rank
+
+
+@dataclass(frozen=True, slots=True)
+class DuplicateLink:
+    duplicate_id: str
+    canonical_id: str
+    kind: str
+    evidence: str
+
+
+def rank(job: Job) -> tuple[int, str]:
+    return source_rank(job.source, job.id)
+
+
+def _ambiguous_urls(jobs: Sequence[Job]) -> frozenset[str]:
+    seen: dict[tuple[str, str], int] = {}
+    for job in jobs:
+        if not job.apply_url_canonical:
+            continue
+        key = (job.source, job.apply_url_canonical)
+        seen[key] = seen.get(key, 0) + 1
+    return frozenset(url for (_, url), count in seen.items() if count > 1)
+
+
+def resolve_duplicates(
+    incoming: Sequence[Job], existing: Sequence[Job]
+) -> tuple[DuplicateLink, ...]:
+    pool: dict[str, Job] = {job.id: job for job in existing}
+    for job in incoming:
+        pool[job.id] = job
+    jobs = sorted(pool.values(), key=rank)
+    ambiguous = _ambiguous_urls(jobs)
+
+    parent: dict[str, str] = {job.id: job.id for job in jobs}
+    boards: dict[str, set[str]] = {job.id: {job.board_key} for job in jobs}
+    sources: dict[str, set[str]] = {job.id: {job.source} for job in jobs}
+    reason: dict[str, tuple[str, str]] = {}
+
+    def find(job_id: str) -> str:
+        while parent[job_id] != job_id:
+            parent[job_id] = parent[parent[job_id]]
+            job_id = parent[job_id]
+        return job_id
+
+    order = {job.id: index for index, job in enumerate(jobs)}
+
+    def union(left_id: str, right_id: str) -> None:
+        left_root, right_root = find(left_id), find(right_id)
+        if left_root == right_root:
+            return
+        if order[right_root] < order[left_root]:
+            left_root, right_root = right_root, left_root
+        parent[right_root] = left_root
+        boards[left_root] |= boards[right_root]
+        sources[left_root] |= sources[right_root]
+
+    for job in jobs:
+        if job.duplicate_of and job.duplicate_of in parent:
+            union(job.id, job.duplicate_of)
+
+    touched = {job.id for job in incoming}
+    for index, left in enumerate(jobs):
+        for right in jobs[index + 1 :]:
+            if left.id not in touched and right.id not in touched:
+                continue
+            left_root, right_root = find(left.id), find(right.id)
+            if left_root == right_root:
+                continue
+            if boards[left_root] & boards[right_root] or sources[left_root] & sources[right_root]:
+                continue
+            match = would_merge(left, right)
+            if not match or (match.kind is MatchKind.URL and left.apply_url_canonical in ambiguous):
+                continue
+            union(left.id, right.id)
+            demoted = right.id if find(right.id) != right_root else left.id
+            reason[demoted] = (match.kind.value, match.evidence)
+
+    links: list[DuplicateLink] = []
+    for job in jobs:
+        canonical = find(job.id)
+        if canonical == job.id:
+            continue
+        kind, evidence = reason.get(job.id, ("", ""))
+        links.append(
+            DuplicateLink(
+                duplicate_id=job.id,
+                canonical_id=canonical,
+                kind=kind,
+                evidence=evidence,
+            )
+        )
+    return tuple(links)