Created image classifier v.1

config/application.yml

@@ -5,3 +5,6 @@ screenshotter:
 
 resolver:
   api_key: 733f6882605be2de8980bbd074091ee4
+
+img_classifier:
+  target_path: "data/screenshots/"

pyproject.toml

@@ -21,7 +21,9 @@ dependencies = [
 dev = [
     "pytest>=9.0.2",
     "ruff==0.15.5",
-    "pytz"
+    "pytz",
+    "types-requests",
+    "types-PyYAML",
     # "playwright==1.58.0" # only dev because it cant be installed in a pipeline, just locally
 ]
 
@@ -37,6 +39,7 @@ lint.select = ["E", "F", "I"]
 python_version = "3.12"
 strict = true
 ignore_missing_imports = true
+plugins = ["pydantic.mypy"]
 
 [tool.pytest.ini_options]
 testpaths = ["test"] 

src/beaky/cli.py

@@ -4,11 +4,10 @@ import yaml
 from pydantic import ValidationError
 
 from beaky.config import Config
+from beaky.link_classifier.classifier import LinkClassifier
+from beaky.resolvers.resolver import _B, _GRAY, _GREEN, _R, _RED, _YELLOW, TicketResolver, TicketVerdict
 from beaky.scanner.scanner import Links
 from beaky.screenshotter.screenshotter import Screenshotter
-from beaky.link_classifier.classifier import LinkClassifier
-from beaky.resolvers.resolver import TicketResolver
-from beaky.resolvers.resolver import TicketVerdict, _R, _B, _GREEN, _RED, _YELLOW, _GRAY
 
 _VERDICT_COLOR = {
     TicketVerdict.TRUTHFUL: _GREEN,

src/beaky/config.py

@@ -1,5 +1,6 @@
 from pydantic.dataclasses import dataclass
 
+from beaky.image_classifier.config import ImgClassifierConfig
 from beaky.resolvers.config import ResolverConfig
 from beaky.screenshotter.config import ScreenshotterConfig
 
@@ -8,4 +9,5 @@ from beaky.screenshotter.config import ScreenshotterConfig
 class Config:
     path: str
     screenshotter: ScreenshotterConfig
-    resolver: ResolverConfig
+    resolver: ResolverConfig
+    img_classifer: ImgClassifierConfig

src/beaky/datamodels/ticket.py

@@ -26,38 +26,38 @@ class Bet(ABC):
     date: datetime
     league: str
     @abstractmethod
-    def resolve(self): pass
+    def resolve(self) -> None: pass
 
 @dataclass
 class WinDrawLose(Bet):
     """Výsledek zápasu 1X2"""
     betType: Literal["X", "0", "1", "2"] = "0"
-    def resolve(self):
+    def resolve(self) -> None:
         ...
 
 @dataclass
 class Advance(Bet):
     """What team advances to next round"""
-    def resolve(self):
+    def resolve(self) -> None:
         raise NotImplementedError("Vyser si voko vine")
 
 @dataclass
 class WinDrawLoseDouble(Bet):
     """Výsledek zápasu - double"""
     betType: Literal["01", "12", "02"] = "01"
-    def resolve(self):
+    def resolve(self) -> None:
         ...
 
 @dataclass
 class WinLose(Bet):
     """Výsledek zápasu bez remízy"""
     betType: Literal["1", "2"] = "1"
-    def resolve(self):
+    def resolve(self) -> None:
         ...
 
 @dataclass
 class BothTeamScored(Bet):
-    def resolve(self):
+    def resolve(self) -> None:
         ...
 
 @dataclass
@@ -65,7 +65,7 @@ class GoalAmount(Bet):
     """Počet gólů v zápasu — over/under total goals"""
     line: float = 0.0      # goal line, e.g. 2.5
     over: bool = True      # True = more than line, False = less than line
-    def resolve(self):
+    def resolve(self) -> None:
         ...
 
 @dataclass
@@ -73,14 +73,14 @@ class GoalHandicap(Bet):
     """Goal handicap for a specific team — add handicap_amount to team's score, team wins = you win"""
     team_bet: Literal["1", "2"] = "1"  # which team the handicap is applied to
     handicap_amount: float = 0.0       # e.g. +1.5 or -0.5
-    def resolve(self):
+    def resolve(self) -> None:
         ...
 
 @dataclass
 class UnknownTicket(Bet):
     """Bet type that could not be classified"""
     raw_text: str = ""
-    def resolve(self):
+    def resolve(self) -> None:
         ...
 
 

src/beaky/image_classifier/classifier.py

@@ -1,7 +1,183 @@
-from datetime import datetime
+import datetime
+from pathlib import Path
 
-from beaky.datamodels.ticket import Ticket
+import pytesseract
+
+from beaky.datamodels.ticket import (
+    Advance,
+    Bet,
+    BetType,
+    BothTeamScored,
+    GoalAmount,
+    GoalHandicap,
+    Ticket,
+    UnknownTicket,
+    WinDrawLose,
+    WinDrawLoseDouble,
+    WinLose,
+)
 
 
-def f(path:str, date:datetime) -> list[Ticket]:
+def img_to_text(path: str) -> str:
-    ...
+    """Given a path to an image, return the text contained in that image.
+    Bypasses PIL and lets Tesseract read the file directly.
+    """
+    try:
+        text = pytesseract.image_to_string(path)
+        return text.strip()
+    except pytesseract.TesseractNotFoundError:
+        print("Error: Tesseract executable not found on your system.")
+        return ""
+    except Exception as e:
+        print(f"Error processing {path}: {e}")
+        return ""
+
+
+def classify(text: str) -> Bet:
+    """Given text extracted from an image and a date, return a Bet object that is
+    relevant to that text."""
+
+    if not text:
+        return UnknownTicket(
+            ticketType=BetType.UNKNOWN,
+            team1Name="N/A",
+            team2Name="N/A",
+            date=datetime.datetime.now(),
+            league="N/A",
+            raw_text="No text extracted",
+        )
+
+    # 1. Defaults & Normalization
+    text_lower = text.lower()
+    date_obj = datetime.datetime.now()
+    team1, team2 = "Unknown", "Unknown"
+    league = "Unknown"
+
+    # 2. Heuristic extraction of Teams (Looking for "Team A - Team B" patterns)
+    lines = [line.strip() for line in text.split("\n") if line.strip()]
+    for line in lines:
+        if " - " in line or " vs " in line or " v " in line:
+            # Avoid splitting on hyphens in dates or numbers
+            if not re.search(r"\d\s*-\s*\d", line):
+                parts = re.split(r" - | vs | v ", line)
+                if len(parts) >= 2:
+                    team1, team2 = parts[0].strip(), parts[1].strip()
+                    break
+
+    # 3. Heuristic extraction of Date (Looking for DD.MM. YYYY HH:MM)
+    date_match = re.search(r"(\d{1,2}\.\s*\d{1,2}\.?\s*(?:\d{2,4})?)\s*(\d{1,2}:\d{2})?", text)
+    if date_match:
+        try:
+            # Fallback to current year if missing, basic parse attempt
+            date_str = f"{date_match.group(1).replace(' ', '')} {date_match.group(2) or '00:00'}"
+            if len(date_str.split(".")[2]) <= 5:  # Missing year
+                date_str = date_str.replace(" ", f"{datetime.datetime.now().year} ")
+            date_obj = datetime.datetime.strptime(date_str, "%d.%m.%Y %H:%M")
+        except Exception:
+            pass  # Keep default if parsing fails
+
+    # 4. Classification Logic based on keywords
+    base_args = {"team1Name": team1, "team2Name": team2, "date": date_obj, "league": league}
+
+    # Advance / Postup
+    if any(kw in text_lower for kw in ["postup", "postoupí", "advance"]):
+        return Advance(ticketType=BetType.ADVANCED, **base_args)
+
+    # Both Teams to Score / Oba dají gól
+    if any(kw in text_lower for kw in ["oba dají gól", "btts", "oba týmy dají gól"]):
+        return BothTeamScored(ticketType=BetType.BOTH_TEAM_SCORED, **base_args)
+
+    # Goal Amount (Over/Under)
+    if any(kw in text_lower for kw in ["počet gólů", "více než", "méně než", "over", "under"]):
+        # Attempt to find the goal line (e.g., 2.5, 3.5)
+        line_match = re.search(r"(\d+\.\d+)", text)
+        line_val = float(line_match.group(1)) if line_match else 2.5
+        is_over = any(kw in text_lower for kw in ["více", "over", "+"])
+
+        return GoalAmount(ticketType=BetType.GOAL_AMOUNT, line=line_val, over=is_over, **base_args)
+
+    # Goal Handicap
+    if any(kw in text_lower for kw in ["handicap", "hcp"]):
+        hcp_match = re.search(r"([+-]?\d+\.\d+)", text)
+        hcp_val = float(hcp_match.group(1)) if hcp_match else 0.0
+        # Simplistic logic: guess team 1 if not explicitly stated
+        team_bet = "2" if " 2 " in text else "1"
+
+        return GoalHandicap(ticketType=BetType.GOAL_HANDICAP, team_bet=team_bet, handicap_amount=hcp_val, **base_args)
+
+    # Win Draw Lose Double (1X, X2, 12)
+    if any(kw in text_lower for kw in ["1x", "x2", "12", "dvojitá šance", "neprohra"]):
+        bet_type = "01" if "1x" in text_lower else "02" if "x2" in text_lower else "12"
+        return WinDrawLoseDouble(ticketType=BetType.WIN_DRAW_LOSE_DOUBLE, betType=bet_type, **base_args)
+
+    # Win Lose (Draw no bet / Vítěz do rozhodnutí)
+    if any(kw in text_lower for kw in ["bez remízy", "vítěz do rozhodnutí", "konečný vítěz"]):
+        bet_type = "2" if re.search(r"\b2\b", text) else "1"
+        return WinLose(ticketType=BetType.WIN_LOSE, betType=bet_type, **base_args)
+
+    # Win Draw Lose (Standard Match Odds)
+    if any(kw in text_lower for kw in ["zápas", "výsledek zápasu", "1x2"]):
+        # Look for isolated 1, X (or 0), or 2
+        match_pick = re.search(r"\b(1|x|0|2)\b", text_lower)
+        bet_type = match_pick.group(1).upper() if match_pick else "1"
+        if bet_type == "X":
+            bet_type = "0"
+
+        return WinDrawLose(ticketType=BetType.WIN_DRAW_LOSE, betType=bet_type, **base_args)
+
+    # Fallback Unknown
+    return UnknownTicket(ticketType=BetType.UNKNOWN, raw_text=text, **base_args)
+
+
+def img_classify(path: str, ticket_id: int) -> Ticket:
+    """Given a path to an image and a date, return a list of Tickets that are
+    relevant to that image and date."""
+    # Define valid image extensions to ignore system files or text documents
+    ticket = Ticket(id=ticket_id, bets=[])
+    valid_extensions = {".png", ".jpg", ".jpeg", ".bmp", ".tiff", ".webp"}
+    path_obj = Path(path)
+
+    if not path_obj.is_dir():
+        print(f"Error: The path '{path}' is not a valid directory.")
+        return ticket
+
+    # Iterate through all files in the folder
+    for file_path in path_obj.iterdir():
+        if file_path.is_file() and file_path.suffix.lower() in valid_extensions:
+            # 1. Extract the text (called separately)
+            extracted_text = img_to_text(str(file_path))
+            print(extracted_text)
+
+            # 2. Classify based on the extracted text (called separately)
+            result = classify(extracted_text)
+
+            # 3. Add the resulting tickets to our main list
+            # Support classifier returning either a single Bet or a list of Bet
+            if result is None:
+                continue
+
+            if isinstance(result, list):
+                for r in result:
+                    print(
+                        r.date,
+                        getattr(r, "ticketType", None),
+                        r.team1Name,
+                        r.team2Name,
+                        r.league,
+                    )
+                ticket.bets.extend(result)
+            else:
+                print(
+                    result.date,
+                    getattr(result, "ticketType", None),
+                    result.team1Name,
+                    result.team2Name,
+                    result.league,
+                )
+                ticket.bets.append(result)
+
+    return ticket
+
+
+if __name__ == "__main__":
+    img_classify("./data/screenshots/", ticket_id=1)

src/beaky/image_classifier/config.py

@@ -0,0 +1,6 @@
+from pydantic.dataclasses import dataclass
+
+
+@dataclass
+class ImgClassifierConfig:
+    target_path: str

src/beaky/link_classifier/classifier.py

@@ -1,19 +1,20 @@
 import re
 from datetime import datetime
+from typing import Any
 
 from playwright.sync_api import Page, sync_playwright
 
 from beaky.datamodels.ticket import (
+    Bet,
+    BetType,
     BothTeamScored,
     GoalAmount,
     GoalHandicap,
     Ticket,
-    BetType,
     UnknownTicket,
     WinDrawLose,
     WinDrawLoseDouble,
     WinLose,
-    Bet
 )
 from beaky.scanner.scanner import Link
 
@@ -37,24 +38,24 @@ def _parse_teams(title: str) -> tuple[str, str]:
 
 
 def _classify_bet(bet_text: str, team1: str, team2: str, date: datetime, league: str) -> Bet:
-    common = dict(team1Name=team1, team2Name=team2, date=date, league=league)
+    common: dict[str, Any] = dict(team1Name=team1, team2Name=team2, date=date, league=league)
 
     # WinDrawLose double: "Výsledek zápasu - dvojtip: 10"
     m = re.search(r"Výsledek zápasu - dvojtip:\s*(\d+)", bet_text)
     if m:
         # normalize order: "10" -> "01", "02" -> "02", "12" -> "12"
         bet_type = "".join(sorted(m.group(1)))
-        return WinDrawLoseDouble(ticketType=BetType.WIN_DRAW_LOSE_DOUBLE, betType=bet_type, **common)
+        return WinDrawLoseDouble(ticketType=BetType.WIN_DRAW_LOSE_DOUBLE, betType=bet_type, **common)  # type: ignore[arg-type]
 
     # WinLose (no draw): "Výsledek bez remízy: 1"
     m = re.search(r"bez rem[ií]zy:\s*([12])", bet_text)
     if m:
-        return WinLose(ticketType=BetType.WIN_LOSE, betType=m.group(1), **common)
+        return WinLose(ticketType=BetType.WIN_LOSE, betType=m.group(1), **common)  # type: ignore[arg-type]
 
     # WinDrawLose: "Výsledek zápasu: 1"
     m = re.search(r"Výsledek zápasu:\s*([012X])\s*$", bet_text.strip())
     if m:
-        return WinDrawLose(ticketType=BetType.WIN_DRAW_LOSE, betType=m.group(1), **common)
+        return WinDrawLose(ticketType=BetType.WIN_DRAW_LOSE, betType=m.group(1), **common)  # type: ignore[arg-type]
 
     # BothTeamScored: "Každý z týmů dá gól v zápasu: Ano"
     if "dá gól" in bet_text or "oba týmy" in bet_text.lower():
@@ -77,7 +78,7 @@ def _classify_bet(bet_text: str, team1: str, team2: str, date: datetime, league:
             return UnknownTicket(ticketType=BetType.UNKNOWN, raw_text=bet_text, **common)
         sign = 1.0 if m.group(1) == "+" else -1.0
         handicap = sign * float(m.group(2))
-        return GoalHandicap(ticketType=BetType.GOAL_HANDICAP, team_bet=team_bet, handicap_amount=handicap, **common)
+        return GoalHandicap(ticketType=BetType.GOAL_HANDICAP, team_bet=team_bet, handicap_amount=handicap, **common)  # type: ignore[arg-type]
 
     return UnknownTicket(ticketType=BetType.UNKNOWN, raw_text=bet_text, **common)
 

src/beaky/resolvers/resolver.py

@@ -1,8 +1,9 @@
 import time
 from dataclasses import dataclass, field
-from datetime import datetime, timedelta
+from datetime import date, datetime, timedelta
 from difflib import SequenceMatcher
 from enum import Enum
+from typing import Any
 
 import requests
 
@@ -149,7 +150,7 @@ class ResolvedTicket:
         return TicketVerdict.TRUTHFUL
 
 
-def _get(url: str, headers: dict, params: dict, retries: int = 3, backoff: float = 60.0) -> requests.Response:
+def _get(url: str, headers: dict[str, str], params: dict[str, str | int], retries: int = 3, backoff: float = 60.0) -> requests.Response:
     for attempt in range(retries):
         resp = requests.get(url, headers=headers, params=params)
         if resp.status_code == 429:
@@ -166,7 +167,7 @@ class TicketResolver:
     def __init__(self, config: ResolverConfig):
         self._headers = {"x-apisports-key": config.api_key}
         # Cache maps (center_date_str, league_id | None) -> list of fixture dicts
-        self._fixture_cache: dict[tuple[str, int | None], list[dict]] = {}
+        self._fixture_cache: dict[tuple[str, int | None], list[dict[str, Any]]] = {}
         # Cache maps league name -> (league_id, confidence)
         self._league_cache: dict[str, tuple[int | None, float]] = {}
 
@@ -215,7 +216,7 @@ class TicketResolver:
             match_finished=finished,
         )
 
-    def _find_fixture(self, bet: Bet) -> tuple[dict | None, float, float, float]:
+    def _find_fixture(self, bet: Bet) -> tuple[dict[str, Any] | None, float, float, float]:
         """Returns (fixture, name_match, date_proximity, league_confidence)."""
         center = bet.date.date()
         date_str = center.strftime("%Y-%m-%d")
@@ -225,7 +226,7 @@ class TicketResolver:
         if cache_key not in self._fixture_cache:
             date_from = (center - timedelta(days=_DATE_WINDOW)).strftime("%Y-%m-%d")
             date_to = (center + timedelta(days=_DATE_WINDOW)).strftime("%Y-%m-%d")
-            params: dict = {"from": date_from, "to": date_to}
+            params: dict[str, str | int] = {"from": date_from, "to": date_to}
             if league_id is not None:
                 params["league"] = league_id
                 params["season"] = center.year if center.month >= 7 else center.year - 1
@@ -273,14 +274,14 @@ def _similarity(a: str, b: str) -> float:
     return SequenceMatcher(None, a.lower(), b.lower()).ratio()
 
 
-def _date_proximity(fixture: dict, center) -> float:
+def _date_proximity(fixture: dict[str, Any], center: date) -> float:
     """1.0 on exact date, linear decay to 0.0 at _DATE_WINDOW days away."""
     fixture_date = datetime.fromisoformat(fixture["fixture"]["date"].replace("Z", "+00:00")).date()
     days_off = abs((fixture_date - center).days)
     return max(0.0, 1.0 - days_off / _DATE_WINDOW)
 
 
-def _best_fixture_match(fixtures: list[dict], team1: str, team2: str, center) -> tuple[dict | None, float, float]:
+def _best_fixture_match(fixtures: list[dict[str, Any]], team1: str, team2: str, center: date) -> tuple[dict[str, Any] | None, float, float]:
     """Returns (best_fixture, name_score, date_proximity) or (None, 0, 0) if no good match."""
     best, best_combined, best_name, best_date = None, 0.0, 0.0, 0.0
     for f in fixtures:
@@ -302,12 +303,12 @@ def _best_fixture_match(fixtures: list[dict], team1: str, team2: str, center) ->
     return (best, best_name, best_date) if best_name > 0.5 else (None, best_name, best_date)
 
 
-def _is_finished(fixture: dict) -> float:
+def _is_finished(fixture: dict[str, Any]) -> float:
     status = fixture.get("fixture", {}).get("status", {}).get("short", "")
     return 1.0 if status in ("FT", "AET", "PEN", "AWD", "WO") else 0.0
 
 
-def _evaluate_bet(bet: Bet, fixture: dict) -> BetOutcome:
+def _evaluate_bet(bet: Bet, fixture: dict[str, Any]) -> BetOutcome:
     goals = fixture.get("goals", {})
     home = goals.get("home")
     away = goals.get("away")

src/beaky/scanner/scanner.py

@@ -21,6 +21,7 @@ class Link:
     url: str
     date: Optional[datetime] = None
 
+
 class Links:
     def __init__(self, path: str | Config):
         if isinstance(path, Config):
@@ -123,10 +124,6 @@ class Links:
         return len(self.links)
 
 
-# Backwards-compatible alias in case other modules referenced Linker
-Linker = Links
-
-
 if __name__ == "__main__":
     links_obj = Links("data/odkazy.xlsx")
     links = links_obj.ret_links()

src/beaky/screenshotter/screenshotter.py

@@ -1,4 +1,5 @@
 from pathlib import Path
+from typing import Any
 
 from playwright.sync_api import sync_playwright
 
@@ -11,7 +12,7 @@ class Screenshotter:
         self.config = config
 
 
-    def capture_tickets(self, links: list[Link]):
+    def capture_tickets(self, links: list[Link]) -> None:
         with sync_playwright() as p:
             browser = p.chromium.launch(headless=True)
             context = browser.new_context()
@@ -24,7 +25,7 @@ class Screenshotter:
 
             browser.close()
 
-    def capture_ticket(self,page, url, target_path, ticket_selector=".betslip-history-detail__left-panel"):
+    def capture_ticket(self, page: Any, url: str, target_path: Path, ticket_selector: str = ".betslip-history-detail__left-panel") -> None:
             page.goto(url)
             page.wait_for_selector(ticket_selector, timeout=10000)
             page.wait_for_timeout(1000)