1# =============================================================================2# Job·Ka — Groupe KA3# Auteur : Simon-Pierre Boucher4# Contact : contact@spboucher.ai5# Fichier : tests/test_normalize.py6# Rôle : Tests unitaires de la couche de normalisation (salaires, dates,7# mode/type d'emploi, lieux, catégories)8# Créé : 2026-08-17 Modifié : 2026-10-209# =============================================================================10import datetime as dt1112from jobka.normalize import (13 categorize,14 clean_html,15 is_quebec_location,16 parse_date,17 parse_employment_type,18 parse_location,19 parse_salary,20 parse_work_mode,21 salary_to_hourly,22 salary_to_yearly,23)2425TODAY = dt.date(2026, 8, 17)262728# --- salaires ----------------------------------------------------------------2930def test_salary_hourly_range():31 assert parse_salary("25,00 $ à 30,00 $ de l'heure") == (25.0, 30.0, "hour")323334def test_salary_hourly_english():35 assert parse_salary("$22.50 - $27.75 per hour") == (22.5, 27.75, "hour")363738def test_salary_yearly_thousands():39 assert parse_salary("70 000 $ à 90 000 $ par année") == (70000.0, 90000.0, "year")404142def test_salary_yearly_english_commas():43 assert parse_salary("$85,000 - $105,000 annually") == (85000.0, 105000.0, "year")444546def test_salary_k_notation():47 assert parse_salary("80k$ - 100k$ /an") == (80000.0, 100000.0, "year")484950def test_salary_single_value():51 lo, hi, unit = parse_salary("Salaire : 28,50 $/h")52 assert (lo, hi, unit) == (28.5, 28.5, "hour")535455def test_salary_infers_unit_from_magnitude():56 assert parse_salary("55 000 $")[2] == "year"57 assert parse_salary("26,75 $")[2] == "hour"585960def test_salary_rejects_hours_of_work():61 # « 35 heures/semaine » n'est PAS un salaire62 assert parse_salary("35 heures par semaine") == (None, None, None)636465def test_salary_rejects_empty_and_garbage():66 assert parse_salary("") == (None, None, None)67 assert parse_salary("Salaire concurrentiel") == (None, None, None)686970def test_salary_conversions():71 assert salary_to_yearly(25.0, "hour") == 52000.072 assert salary_to_hourly(52000.0, "year") == 25.073 assert salary_to_yearly(5000.0, "month") == 60000.074 assert salary_to_yearly(None, "hour") is None757677# --- dates ---------------------------------------------------------------------7879def test_date_iso_passthrough():80 assert parse_date("2026-09-01") == "2026-09-01"81 assert parse_date("2026-06-16T05:06:55-04:00") == "2026-06-16"828384def test_date_epoch_ms():85 # Lever : createdAt en millisecondes86 assert parse_date(1755000000000) == dt.date.fromtimestamp(1755000000).isoformat()878889def test_date_french():90 assert parse_date("1er juillet 2026", TODAY) == "2026-07-01"91 assert parse_date("déc. 2026", TODAY) == "2026-12-01"929394def test_date_english():95 assert parse_date("July 1st, 2026", TODAY) == "2026-07-01"969798def test_date_java_tostring():99 # SuccessFactors CSB : meta itemprop="datePosted" au format100 # java.util.Date.toString() (cas Dentons careers.dentons.com, 2026-10-06)101 assert parse_date("Sun Oct 04 02:00:00 UTC 2026") == "2026-10-04"102 assert parse_date("Mon Dec 01 23:00:00 EST 2026") == "2026-12-01"103104105def test_date_relative():106 assert parse_date("Posted 3 Days Ago", TODAY) == "2026-08-14"107 assert parse_date("il y a 2 semaines", TODAY) == "2026-08-03"108 assert parse_date("Posted Today", TODAY) == "2026-08-17"109 assert parse_date("Posted Yesterday", TODAY) == "2026-08-16"110111112def test_date_never_invented():113 assert parse_date("bientôt") is None114 assert parse_date("") is None115 assert parse_date(None) is None116117118# --- mode / type ---------------------------------------------------------------119120def test_work_mode():121 assert parse_work_mode("Télétravail") == "teletravail"122 assert parse_work_mode("Remote - QC") == "teletravail"123 assert parse_work_mode("Mode hybride (3 jours au bureau)") == "hybride"124 assert parse_work_mode("Sur place") == "presentiel"125 assert parse_work_mode("Montréal") is None126127128def test_work_mode_hybrid_wins():129 assert parse_work_mode("hybride avec télétravail partiel") == "hybride"130131132def test_employment_type():133 assert parse_employment_type("Temps plein") == "temps_plein"134 assert parse_employment_type("Full time") == "temps_plein"135 assert parse_employment_type("Permanent") == "temps_plein"136 assert parse_employment_type("Temps partiel") == "temps_partiel"137 assert parse_employment_type("Contrat 12 mois") == "contractuel"138 assert parse_employment_type("Stage été 2027") == "stage"139 assert parse_employment_type("Emploi saisonnier") == "saisonnier"140 assert parse_employment_type("") is None141142143# --- lieux -----------------------------------------------------------------------144145def test_location_montreal():146 loc = parse_location("Montréal, QC, Canada")147 assert loc["city"] == "Montréal"148 assert loc["region"] == "Québec"149150151def test_location_quebec_city_english():152 assert parse_location("Quebec City, Quebec")["city"] == "Québec"153154155def test_location_postal_code():156 loc = parse_location("1500 rue Exemple, Lévis G6V 4Z2")157 assert loc["postal_code"] == "G6V 4Z2"158 assert loc["city"] == "Lévis"159160161def test_is_quebec_location():162 assert is_quebec_location("Montréal, QC")163 assert is_quebec_location("Saint-Laurent, Quebec, Canada")164 assert is_quebec_location("Sherbrooke")165 # « Sitec Fermont » : site minier de Cambria à Fermont (Côte-Nord) —166 # libellé Workday sans province, la ville seule doit suffire (bug cambria167 # 2026-09-12 : Fermont absent de _QC_CITIES rendait l'offre invisible)168 assert is_quebec_location("Sitec Fermont")169 # « Baie-D'Urfé » : libellé Workday nu sans province (bug galderma170 # 2026-10-20 : le tenant a relibellé son usine « Montreal » ->171 # « Baie-D'Urfé » avec ids republiés — ville absente de _QC_CITIES,172 # found 16->0 alors que le board portait 49 offres QC vivantes)173 assert is_quebec_location("Baie-D'Urfé")174 assert is_quebec_location("Baie D'Urfe, QC")175 assert not is_quebec_location("Toronto, ON")176 assert not is_quebec_location("Richmond, BC")177 assert not is_quebec_location("Paris, IDF, France")178 assert not is_quebec_location("")179 # « Quebec St./Street/Ave » = nom de RUE hors Québec, pas la province (bug180 # solmentalhealth 2026-09-14 : bureau de Denver sur Quebec Street publié181 # comme Québec/Capitale-Nationale — le marqueur \bquebec\b gagnait avant182 # le rejet « , CO »)183 assert not is_quebec_location("3401 Quebec St. Suite 8000, Denver, CO, 80207")184 assert not is_quebec_location("3401 Quebec St. Suite 8000")185 assert not is_quebec_location("2600 Quebec Street NW, Washington, DC")186 assert not is_quebec_location("Quebec Ave, Golden Valley, MN")187 # …mais les abréviations de saints des libellés QC légitimes survivent188 assert is_quebec_location("CAN Quebec St-Hubert")189 assert is_quebec_location("CAN Quebec Ste-Thérèse")190 assert is_quebec_location("CAN Quebec Plessisville")191 assert not parse_location("3401 Quebec St. Suite 8000, Denver, CO, 80207")["region"]192193194# --- catégories / HTML ------------------------------------------------------------195196def test_categorize():197 assert categorize("Développeur logiciel senior") == "TI"198 assert categorize("Infirmière clinicienne") == "Santé"199 assert categorize("Conseiller en ressources humaines") == "RH"200 assert categorize("Poste mystère") == ""201202203def test_clean_html():204 html = "<p>Bonjour !</p><ul><li>Point 1</li><li>Point 2</li></ul>"205 txt = clean_html(html)206 assert "Bonjour !" in txt207 assert "Point 1" in txt and "Point 2" in txt208 assert "<" not in txt209210211# --- Phase 2 (2026-08-19) : enrichissement ------------------------------------212213def test_clean_html_double_escaped():214 # contenu sur-échappé (Greenhouse/Lever) : entités dans les entités215 html = "<p>Bonjour&nbsp;!</p><li>Point</li>"216 txt = clean_html(html)217 assert "<" not in txt and " " not in txt218 assert "Bonjour" in txt and "Point" in txt219220221def test_sanitize_salary_month_is_actually_year():222 from jobka.normalize import sanitize_salary223 # « 75 000 $/mois » : hors bornes mensuelles -> réinterprété en annuel224 assert sanitize_salary(75000, 82500, "month") == (75000, 82500, "year")225226227def test_sanitize_salary_hour_is_actually_year():228 from jobka.normalize import sanitize_salary229 assert sanitize_salary(57200, None, "hour") == (57200, None, "year")230231232def test_sanitize_salary_plausible_untouched():233 from jobka.normalize import sanitize_salary234 assert sanitize_salary(25.0, 30.0, "hour") == (25.0, 30.0, "hour")235236237def test_sanitize_salary_absurd_dropped():238 from jobka.normalize import sanitize_salary239 assert sanitize_salary(2.0, 3.0, "hour") == (None, None, None)240241242def test_salary_from_text_skips_funding():243 from jobka.normalize import salary_from_text244 lo, _, _, _ = salary_from_text(245 "Nous offrons 600 000 $ en financement de pré-amorçage aux fondateurs.")246 assert lo is None247248249def test_salary_from_text_decimal_amounts():250 from jobka.normalize import salary_from_text251 lo, hi, unit, _ = salary_from_text(252 "The annual salary range is $82,500.00 to $137,500.00 per year.")253 assert (lo, hi, unit) == (82500.0, 137500.0, "year")254255256def test_salary_from_text_skips_expense():257 from jobka.normalize import salary_from_text258 lo, _, _, _ = salary_from_text(259 "You get a 250$ BEHappy expense reimbursement to buy what you want.")260 assert lo is None261262263def test_detect_language_fr():264 from jobka.normalize import detect_language265 assert detect_language(266 "Développeur logiciel",267 "Nous cherchons une personne pour rejoindre notre équipe dans nos "268 "bureaux de Montréal. Le poste est pour vous si vous aimez le défi "269 "et les tâches variées avec des exigences claires.") == "fr"270271272def test_detect_language_en():273 from jobka.normalize import detect_language274 assert detect_language(275 "Software Developer",276 "We are looking for someone to join our team in Montreal. You will "277 "work with the best and your skills will grow as we build this "278 "product for the world.") == "en"279280281def test_detect_language_too_short():282 from jobka.normalize import detect_language283 assert detect_language("Poste", "") == ""284285286def test_extract_benefits_block():287 from jobka.normalize import extract_benefits288 desc = ("Description du poste.\n\nAvantages\n"289 "• Assurances collectives\n• REER avec cotisation de l'employeur\n"290 "• 4 semaines de vacances\n\nComment postuler\nEnvoyez votre CV.")291 ben = extract_benefits(desc)292 assert "Assurances collectives" in ben293 assert "4 semaines de vacances" in ben294 assert len(ben) == 3295296297def test_extract_benefits_none_without_block():298 from jobka.normalize import extract_benefits299 assert extract_benefits("Un poste intéressant sans liste structurée.") == []300301302def test_clean_title_requisition_code():303 from jobka.normalize import clean_title304 assert clean_title("Analyste financier - R-12345") == "Analyste financier"305 assert clean_title("[REQ-9876] Technicien réseau") == "Technicien réseau"306307308def test_clean_title_city_suffix():309 from jobka.normalize import clean_title310 assert clean_title("Développeur Python - Montréal", "Montréal") == \311 "Développeur Python"312 # ville différente : suffixe conservé313 assert clean_title("Développeur Python - Montréal", "Laval") == \314 "Développeur Python - Montréal"315316317def test_clean_title_all_caps():318 from jobka.normalize import clean_title319 t = clean_title("CONSEILLER RH ET ACQUISITION DE TALENTS")320 assert t != t.upper()321 assert "RH" in t322323324def test_region_for_city():325 from jobka.regions import region_for_city326 assert region_for_city("Montréal") == "Montréal"327 assert region_for_city("Québec") == "Capitale-Nationale"328 assert region_for_city("Lévis") == "Chaudière-Appalaches"329 assert region_for_city("Dorval") == "Montréal"330 assert region_for_city("Ville inconnue XYZ") == ""331332333def test_quality_issues():334 from jobka.quality import issues335 ok = issues(title="Développeur", description="x" * 100, city="Montréal",336 date_posted="2026-08-01", date_deadline="2026-09-01",337 salary_hour_min=25.0, salary_hour_max=30.0,338 salary_year_min=None, salary_year_max=None,339 salary_unit="hour")340 assert ok == []341 bad = issues(title="", description="court", city="Toronto",342 date_posted="2026-08-01", date_deadline="2026-07-01",343 salary_hour_min=300.0, salary_hour_max=None,344 salary_year_min=None, salary_year_max=None,345 salary_unit="hour")346 assert set(bad) == {"titre_vide", "description_courte", "ville_hors_quebec",347 "dates_incoherentes", "salaire_hors_bornes"}348349350def test_quality_high_yearly_not_hourly_capped():351 from jobka.quality import issues352 # 457 500 $/an (~ 220 $/h equiv.) : legitime, sous la borne annuelle 500 k$353 assert issues(title="Directeur", description="x" * 100, city="Montréal",354 date_posted=None, date_deadline=None,355 salary_hour_min=131.97, salary_hour_max=219.95,356 salary_year_min=274500.0, salary_year_max=457500.0,357 salary_unit="year") == []358