@camstack/addon-waste-collections 0.0.0-stage → 0.1.2

package/python/vendor/waste_collection_schedule/source/smolo_cz.py

@@ -0,0 +1,566 @@
+"""Source for SMOLO a.s. (Třinec), Czech Republic."""
+
+from __future__ import annotations
+
+import itertools
+import logging
+import re
+import unicodedata
+import urllib.parse
+from dataclasses import dataclass, field
+from datetime import date, timedelta
+from io import BytesIO
+
+import requests
+from bs4 import BeautifulSoup
+from pypdf import PdfReader
+from waste_collection_schedule import Collection, Icons  # type: ignore[attr-defined]
+from waste_collection_schedule.exceptions import SourceArgumentNotFoundWithSuggestions
+
+_LOGGER = logging.getLogger(__name__)
+
+TITLE = "SMOLO a.s. (Třinec)"
+DESCRIPTION = (
+    "Source for SMOLO a.s. household waste collection schedule for the town "
+    "of Třinec, Czech Republic."
+)
+URL = "https://www.smolo.cz/sluzby/odpadove-hospodarstvi/informace-pro-domacnosti/"
+COUNTRY = "cz"
+
+SOURCE_CODEOWNERS = ["@jan-tdy"]
+
+_LANDING_PAGE_URL = (
+    "https://www.smolo.cz/sluzby/odpadove-hospodarstvi/informace-pro-domacnosti/"
+)
+
+_WEEKDAY_TOKENS = {"PO": 0, "ÚT": 1, "ST": 2, "ČT": 3, "PÁ": 4}  # Monday=0 .. Friday=4
+
+_SKO_T = "Směsný komunální odpad"
+_BIO_T = "Biologický odpad"
+_SEP_T = "Tříděný odpad (pytle/240L nádoby)"
+_SEP_YELLOW_T = "Tříděný odpad - plast (žluté pytle/nádoby)"
+_SEP_BLUE_T = "Tříděný odpad - papír (modré pytle/nádoby)"
+
+ICON_MAP = {
+    _SKO_T: Icons.GENERAL_WASTE,
+    _BIO_T: Icons.ORGANIC,
+    _SEP_T: Icons.RECYCLING,
+    _SEP_YELLOW_T: Icons.RECYCLING,
+    _SEP_BLUE_T: Icons.PAPER,
+}
+
+TEST_CASES = {
+    "Nebory": {"district": "Nebory"},
+    "Karpentná": {"district": "Karpentná"},
+    "Stare Mesto Husova": {"district": "Staré Město", "street": "Husova"},
+}
+
+HOW_TO_GET_ARGUMENTS_DESCRIPTION = {
+    "en": (
+        "Open the SMOLO a.s. household-waste page "
+        "(https://www.smolo.cz/sluzby/odpadove-hospodarstvi/informace-pro-domacnosti/) "
+        "and look up the current 'Třinec – harmonogram' PDF for the SKO/BIO/SEP "
+        "schedule. Use the 'ČÁST MĚSTA' (part of town / district) column value "
+        "as the 'district' argument, e.g. 'Staré Město', 'Lyžbice', 'Nebory', "
+        "'Karpentná'. If your district appears multiple times in the table "
+        "with different street lists, also set 'street' to one street name "
+        "from the 'LOKALITA' column to disambiguate."
+    ),
+}
+
+PARAM_TRANSLATIONS = {
+    "en": {
+        "district": "District (část města)",
+        "street": "Street (optional, for disambiguation)",
+    },
+}
+PARAM_DESCRIPTIONS = {
+    "en": {
+        "district": (
+            "The town district / part of town ('část města') exactly as it "
+            "appears in the SMOLO Třinec collection schedule, e.g. 'Staré "
+            "Město', 'Lyžbice', 'Nebory', 'Karpentná'."
+        ),
+        "street": (
+            "Optional street name used to disambiguate districts that appear "
+            "on multiple rows of the schedule with different street lists "
+            "(e.g. 'Staré Město', 'Oldřichovice', 'Lyžbice')."
+        ),
+    },
+}
+
+
+def _normalize(text: str) -> str:
+    """Lowercase and strip diacritics for tolerant matching."""
+    normalized = unicodedata.normalize("NFKD", text)
+    without_diacritics = "".join(c for c in normalized if not unicodedata.combining(c))
+    return without_diacritics.lower().strip()
+
+
+def _fields_of(line: str) -> list[tuple]:
+    """Split a layout-mode-extracted line into (start_col, text) fields,
+    where fields are separated by runs of 2+ spaces."""
+    return [(m.start(), m.group()) for m in re.finditer(r"\S(?:[^\s]|\s(?!\s))*", line)]
+
+
+def _find_split_column(starts: list[int]) -> int:
+    """Find the largest gap in a sorted list of column-start positions to
+    split fields into a 'district' cluster and a 'street' cluster."""
+    unique_starts = sorted(set(starts))
+    if len(unique_starts) < 2:
+        return (unique_starts[0] + 1) if unique_starts else 20
+    best_gap = -1
+    split_col = unique_starts[0] + 1
+    for a, b in itertools.pairwise(unique_starts):
+        if b - a > best_gap:
+            best_gap = b - a
+            split_col = (a + b) // 2
+    return split_col
+
+
+@dataclass
+class _Page1Row:
+    weekday: int
+    district_raw: str
+    streets: list[str] = field(default_factory=list)
+
+    @property
+    def bio_on_monday(self) -> bool:
+        return "bio*" in _normalize(self.district_raw)
+
+    @property
+    def district_names(self) -> list[str]:
+        # Strip any "(...)" annotation, then split on commas for
+        # multi-district cells like "Terasa, Lyžbice Ves".
+        cleaned = re.sub(r"\(.*?\)", "", self.district_raw)
+        return [p.strip() for p in cleaned.split(",") if p.strip()]
+
+    @property
+    def street_text(self) -> str:
+        return " ".join(self.streets)
+
+
+def _parse_page1(text: str) -> tuple:
+    """Parse the SKO/BIO page (page 1) of the schedule PDF.
+
+    Returns (rows, prose_lines).
+    """
+    lines = list(text.split("\n"))
+    header_idx = next(
+        (
+            i
+            for i, line_ in enumerate(lines)
+            if "ČÁST MĚSTA" in line_ and "LOKALITA" in line_
+        ),
+        None,
+    )
+    if header_idx is None:
+        return [], []
+
+    # The header wraps ("SVOZOVÝ" / "DEN"); the table body starts two lines down.
+    body_start = header_idx + 2
+
+    body_end = len(lines)
+    for i in range(body_start, len(lines)):
+        if lines[i] and not lines[i][0].isspace():
+            body_end = i
+            break
+
+    body_lines = lines[body_start:body_end]
+
+    non_weekday_starts = [
+        s
+        for line_ in body_lines
+        for s, t in _fields_of(line_)
+        if t not in _WEEKDAY_TOKENS
+    ]
+    split_col = _find_split_column(non_weekday_starts)
+
+    # Pre-scan for the first weekday token: it may appear visually mid-block
+    # (vertically centered within its block of rows) rather than on the very
+    # first row of that block, but the first block in the table always starts
+    # at the top regardless of where its token lands.
+    current_weekday = None
+    for line_ in body_lines:
+        for _, t in _fields_of(line_):
+            if t in _WEEKDAY_TOKENS:
+                current_weekday = _WEEKDAY_TOKENS[t]
+                break
+        if current_weekday is not None:
+            break
+
+    rows: list[_Page1Row] = []
+    open_row: _Page1Row | None = None
+    pending_before: list[str] = []
+
+    for line_ in body_lines:
+        dist_field = None
+        street_fields = []
+        for s, t in _fields_of(line_):
+            if t in _WEEKDAY_TOKENS:
+                current_weekday = _WEEKDAY_TOKENS[t]
+                continue
+            if s < split_col:
+                dist_field = t
+            else:
+                street_fields.append(t)
+        street_text = ", ".join(street_fields)
+
+        if dist_field and not dist_field.startswith("("):
+            row = _Page1Row(
+                weekday=current_weekday if current_weekday is not None else 0,
+                district_raw=dist_field,
+                streets=list(pending_before),
+            )
+            pending_before = []
+            if street_text:
+                row.streets.append(street_text)
+            rows.append(row)
+            open_row = row
+        elif dist_field and dist_field.startswith("("):
+            # continuation of the open row's district annotation, e.g.
+            # "(BIO* svoz v pondělí)" wrapping onto its own line
+            if open_row is not None:
+                open_row.district_raw += " " + dist_field
+                if street_text:
+                    open_row.streets.append(street_text)
+            elif street_text:
+                pending_before.append(street_text)
+        else:
+            if street_text:
+                if open_row is not None:
+                    open_row.streets.append(street_text)
+                else:
+                    pending_before.append(street_text)
+
+    return rows, lines[body_end:]
+
+
+_DATE_FIELD_RE = re.compile(r"\d{1,2}[.,]\d{1,2}\.?")
+_WD_FIELD_RE = re.compile(r"^(PO|ÚT|ST|ČT|PÁ)(?:\s?(\d.*))?$")
+
+
+@dataclass
+class _Page2Block:
+    weekday: int
+    dates: list[str]
+    text: list[str]
+
+    @property
+    def joined_text(self) -> str:
+        return " ".join(self.text)
+
+
+def _parse_page2(text: str) -> list[_Page2Block]:
+    """Parse the SEP (sorted waste) page (page 2) of the schedule PDF."""
+    lines = list(text.split("\n"))
+    header_idx = next(
+        (i for i, line_ in enumerate(lines) if "Část" in line_ and "lokality" in line_),
+        None,
+    )
+    if header_idx is None:
+        return []
+
+    # The header wraps onto a continuation line; skip to the first blank line.
+    body_start = header_idx + 1
+    while body_start < len(lines) and lines[body_start].strip():
+        body_start += 1
+
+    blocks: list[_Page2Block] = []
+    cur_text: list[str] = []
+    cur_weekday: int | None = None
+    cur_dates: list[str] = []
+
+    def flush() -> None:
+        nonlocal cur_text, cur_weekday, cur_dates
+        if cur_weekday is not None and cur_text:
+            blocks.append(
+                _Page2Block(
+                    weekday=cur_weekday, dates=list(cur_dates), text=list(cur_text)
+                )
+            )
+        cur_text, cur_weekday, cur_dates = [], None, []
+
+    for line_ in lines[body_start:]:
+        if not line_.strip():
+            if cur_weekday is not None:
+                flush()
+            continue
+
+        line_text: list[str] = []
+        line_weekday: int | None = None
+        line_dates: list[str] = []
+        for _, tok in _fields_of(line_):
+            m = _WD_FIELD_RE.match(tok)
+            if m:
+                line_weekday = _WEEKDAY_TOKENS[m.group(1)]
+                if m.group(2):
+                    line_dates.extend(_DATE_FIELD_RE.findall(m.group(2)))
+                continue
+            if re.fullmatch(r"[\d.,\s]+", tok):
+                found = _DATE_FIELD_RE.findall(tok)
+                if found:
+                    line_dates.extend(found)
+                    continue
+            line_text.append(tok)
+
+        if line_weekday is not None and cur_weekday is not None:
+            flush()
+
+        cur_text.extend(line_text)
+        if line_weekday is not None:
+            cur_weekday = line_weekday
+        cur_dates.extend(line_dates)
+
+    flush()
+    return blocks
+
+
+_PLATNY_OD_RE = re.compile(r"PLATN[ÝY]\s+OD\s+(\d{1,2})\.(\d{1,2})\.(\d{4})")
+_HOLIDAY_SHIFT_RE = re.compile(
+    r"(\d{1,2}\.\d{1,2}\.\d{4})\s*\(svoz se\s*posouvá na \S+\s+(\d{1,2}\.\d{1,2}\.\d{4})\)"
+)
+_BIO_WEEKS_RE = re.compile(r"\(([\d.,\s]+)týden roku (\d{4})\)")
+
+
+def _parse_year(text: str) -> int | None:
+    m = _PLATNY_OD_RE.search(text)
+    if m:
+        return int(m.group(3))
+    return None
+
+
+def _parse_date_ddmmyyyy(text: str) -> date | None:
+    parts = text.split(".")
+    try:
+        day, month, year = int(parts[0]), int(parts[1]), int(parts[2])
+        return date(year, month, day)
+    except (ValueError, IndexError):
+        return None
+
+
+def _parse_holiday_shifts(prose_lines: list[str]) -> dict:
+    joined = " ".join(prose_lines)
+    shifts = {}
+    for m in _HOLIDAY_SHIFT_RE.finditer(joined):
+        from_date = _parse_date_ddmmyyyy(m.group(1))
+        to_date = _parse_date_ddmmyyyy(m.group(2))
+        if from_date and to_date:
+            shifts[from_date] = to_date
+    if not shifts:
+        _LOGGER.warning(
+            "smolo_cz: could not parse holiday-shift sentence from the SKO/BIO "
+            "PDF text; holiday shifts will not be applied this run."
+        )
+    return shifts
+
+
+def _parse_sparse_bio_weeks(prose_lines: list[str]) -> set | None:
+    joined = " ".join(prose_lines)
+    m = _BIO_WEEKS_RE.search(joined)
+    if not m:
+        _LOGGER.warning(
+            "smolo_cz: could not parse the 'BIO 1x/month in Jan/Feb/Dec' week "
+            "numbers from the PDF text; falling back to 'every even week in "
+            "those months' as a simplification."
+        )
+        return None
+    weeks = set()
+    for chunk in m.group(1).split(","):
+        chunk = chunk.strip().rstrip(".")
+        if chunk.isdigit():
+            weeks.add(int(chunk))
+    return weeks or None
+
+
+def _iter_year_dates(year: int):
+    d = date(year, 1, 1)
+    end = date(year, 12, 31)
+    while d <= end:
+        yield d
+        d += timedelta(days=1)
+
+
+class Source:
+    def __init__(self, district: str, street: str = "") -> None:
+        self._district = district
+        self._street = street
+
+    def _fetch_pdf_url(self, session: requests.Session) -> str:
+        response = session.get(_LANDING_PAGE_URL, timeout=30)
+        response.raise_for_status()
+        soup = BeautifulSoup(response.text, "html.parser")
+
+        for a in soup.find_all("a", href=True):
+            href = str(a["href"])
+            if ".pdf" not in href.lower():
+                continue
+            link_descriptor = _normalize(f"{a.get_text(' ', strip=True)} {href}")
+            if "trinec" in link_descriptor and all(
+                token in link_descriptor for token in ("sko", "bio", "sep")
+            ):
+                return urllib.parse.urljoin(_LANDING_PAGE_URL, href)
+
+        raise SourceArgumentNotFoundWithSuggestions(
+            "district",
+            self._district,
+            [
+                "Could not locate the Třinec SKO/BIO/SEP schedule PDF link on "
+                "the SMOLO household-waste page. The page layout may have "
+                "changed."
+            ],
+        )
+
+    def _match_page1_rows(self, rows: list[_Page1Row]) -> list[_Page1Row]:
+        target_district = _normalize(self._district)
+        target_street = _normalize(self._street) if self._street else ""
+
+        district_matches = [
+            row
+            for row in rows
+            if any(_normalize(n) == target_district for n in row.district_names)
+        ]
+
+        if not district_matches:
+            all_names = sorted({n for row in rows for n in row.district_names})
+            raise SourceArgumentNotFoundWithSuggestions(
+                "district", self._district, all_names
+            )
+
+        if len(district_matches) == 1:
+            return district_matches
+
+        if target_street:
+            street_matches = [
+                row
+                for row in district_matches
+                if target_street in _normalize(row.street_text)
+            ]
+            if len(street_matches) == 1:
+                return street_matches
+            if street_matches:
+                # multiple rows still match (unlikely) -- return them all
+                return street_matches
+
+        suggestions = [
+            f"{row.district_raw} ({row.street_text})" for row in district_matches
+        ]
+        raise SourceArgumentNotFoundWithSuggestions(
+            "street" if target_street else "district",
+            self._street if target_street else self._district,
+            suggestions,
+        )
+
+    def _match_page2_blocks(self, blocks: list[_Page2Block]) -> list[_Page2Block]:
+        target_district = _normalize(self._district)
+        matches = [
+            block
+            for block in blocks
+            if target_district in _normalize(block.joined_text)
+        ]
+        # The Nebory yellow/plastic (žlutá) and blue/paper (modrá) rows only
+        # mention "Nebory" once (vertically centered between the two rows in
+        # the source PDF); attach both explicitly when the district is Nebory.
+        if target_district == "nebory":
+            for block in blocks:
+                text_norm = _normalize(block.joined_text)
+                if (
+                    "nadoba" in text_norm or "pytle" in text_norm
+                ) and block not in matches:
+                    if "zluta" in text_norm or "modra" in text_norm:
+                        matches.append(block)
+        return matches
+
+    def fetch(self) -> list[Collection]:
+        session = requests.Session()
+        session.headers.update(
+            {"User-Agent": "Mozilla/5.0 (compatible; HomeAssistant)"}
+        )
+
+        pdf_url = self._fetch_pdf_url(session)
+        pdf_response = session.get(pdf_url, timeout=30)
+        pdf_response.raise_for_status()
+
+        reader = PdfReader(BytesIO(pdf_response.content))
+        page1_text = reader.pages[0].extract_text(extraction_mode="layout") or ""
+        page2_text = (
+            reader.pages[1].extract_text(extraction_mode="layout")
+            if len(reader.pages) > 1
+            else ""
+        ) or ""
+
+        rows, prose_lines = _parse_page1(page1_text)
+        if not rows:
+            raise SourceArgumentNotFoundWithSuggestions(
+                "district",
+                self._district,
+                ["Could not parse the SKO/BIO schedule table from the PDF."],
+            )
+
+        year = _parse_year(page1_text)
+        if year is None:
+            raise ValueError("Could not parse the schedule year from the SKO/BIO PDF.")
+        holiday_shifts = _parse_holiday_shifts(prose_lines)
+        sparse_bio_weeks = _parse_sparse_bio_weeks(prose_lines)
+
+        matched_rows = self._match_page1_rows(rows)
+
+        entries: list[Collection] = []
+
+        for row in matched_rows:
+            bio_weekday = 0 if row.bio_on_monday else row.weekday
+
+            for d in _iter_year_dates(year):
+                iso_week = d.isocalendar()[1]
+                is_odd_week = iso_week % 2 == 1
+
+                if d.weekday() == row.weekday and is_odd_week:
+                    final_date = holiday_shifts.get(d, d)
+                    entries.append(
+                        Collection(date=final_date, t=_SKO_T, icon=ICON_MAP[_SKO_T])
+                    )
+
+                if d.weekday() == bio_weekday and not is_odd_week:
+                    if d.month in (1, 2, 12):
+                        if sparse_bio_weeks is not None:
+                            if iso_week not in sparse_bio_weeks:
+                                continue
+                        # else: fall back to "every even week" simplification
+                    final_date = holiday_shifts.get(d, d)
+                    entries.append(
+                        Collection(date=final_date, t=_BIO_T, icon=ICON_MAP[_BIO_T])
+                    )
+
+        # SEP (sorted waste) — page 2, explicit date lists.
+        page2_blocks = _parse_page2(page2_text)
+        page2_year = _parse_year(page2_text) or year
+        matched_blocks = self._match_page2_blocks(page2_blocks)
+
+        for block in matched_blocks:
+            text_norm = _normalize(block.joined_text)
+            if "zluta" in text_norm:
+                sep_t = _SEP_YELLOW_T
+            elif "modra" in text_norm:
+                sep_t = _SEP_BLUE_T
+            else:
+                sep_t = _SEP_T
+
+            for raw_date in block.dates:
+                parts = re.split(r"[.,]", raw_date)
+                parts = [p for p in parts if p]
+                if len(parts) < 2:
+                    _LOGGER.warning(
+                        "smolo_cz: skipping unparsable SEP date fragment %r", raw_date
+                    )
+                    continue
+                try:
+                    day, month = int(parts[0]), int(parts[1])
+                    d = date(page2_year, month, day)
+                except ValueError:
+                    _LOGGER.warning(
+                        "smolo_cz: skipping unparsable SEP date fragment %r", raw_date
+                    )
+                    continue
+                entries.append(Collection(date=d, t=sep_t, icon=ICON_MAP[sep_t]))
+
+        return entries

package/python/vendor/waste_collection_schedule/source/snaga_mb_si.py

@@ -0,0 +1,72 @@
+from datetime import date, datetime
+
+import requests
+from waste_collection_schedule import Collection, Icons  # type: ignore[attr-defined]
+from waste_collection_schedule.exceptions import SourceArgumentExceptionMultiple
+
+TITLE = "Snaga Maribor"
+DESCRIPTION = "Source for Snaga Maribor."
+URL = "https://snaga-mb.si/"
+TEST_CASES = {
+    "Ruska ulica 24": {"street": "Ruska ulica", "house_number": 24},
+}
+
+
+BIN_TYPES = {
+    "O": "Ost.",
+    "B": "Bio.",
+    "U": "Emb.",
+    "P": "Pap.",
+    "S": "Stek.",
+}
+ICON_MAP = {
+    "O": Icons.GENERAL_WASTE,
+    "B": Icons.ORGANIC,
+    "U": Icons.RECYCLING,
+    "P": Icons.NEWSPAPER,
+    "S": Icons.GLASS,
+}
+
+BASE_URL = "https://arhiv.snaga-mb.si/mso"
+ADDRESS_ID_URL = f"{BASE_URL}/tmRSkjpgG.php"
+COLLECTION_URL = f"{BASE_URL}/tmRSk.php"
+
+
+class Source:
+    def __init__(self, street: str, house_number: str | int):
+        self._street: str = street
+        self._house_number: str | int = house_number
+
+    def fetch(self) -> list[Collection]:
+        args = {
+            "Naziv": self._street,
+            "HS": self._house_number,
+            "IDo": "0",
+            "_": int(datetime.now().timestamp() * 1000),
+        }
+        r = requests.get(ADDRESS_ID_URL, params=args)
+        r.raise_for_status()
+        if r.text == "null" or not r.json():
+            raise SourceArgumentExceptionMultiple(
+                ["street", "house_number"], "Invalid address"
+            )
+
+        entries = []
+        for id in r.json():
+            address_id = id["OM"]
+            r = requests.get(
+                COLLECTION_URL,
+                params={"sql": address_id, "_": int(datetime.now().timestamp() * 1000)},
+            )
+
+            for collection in r.json():
+                day = int(collection["Dan"])
+                month = int(collection["Mesec"])
+                year = int(collection["Leto"])
+                date_ = date(year, month, day)
+                type_char = collection["Odp"]
+                bin_type = BIN_TYPES.get(type_char, type_char)
+                icon = ICON_MAP.get(type_char)
+                entries.append(Collection(date=date_, t=bin_type, icon=icon))
+
+        return entries