Initial commit — PlantGuideScraper project

backend/app/schemas/__init__.py

@@ -0,0 +1,15 @@
+from app.schemas.species import SpeciesCreate, SpeciesUpdate, SpeciesResponse, SpeciesListResponse
+from app.schemas.image import ImageResponse, ImageListResponse, ImageFilter
+from app.schemas.job import JobCreate, JobResponse, JobListResponse
+from app.schemas.api_key import ApiKeyCreate, ApiKeyUpdate, ApiKeyResponse
+from app.schemas.export import ExportCreate, ExportResponse, ExportListResponse
+from app.schemas.stats import StatsResponse, SourceStats, SpeciesStats
+
+__all__ = [
+    "SpeciesCreate", "SpeciesUpdate", "SpeciesResponse", "SpeciesListResponse",
+    "ImageResponse", "ImageListResponse", "ImageFilter",
+    "JobCreate", "JobResponse", "JobListResponse",
+    "ApiKeyCreate", "ApiKeyUpdate", "ApiKeyResponse",
+    "ExportCreate", "ExportResponse", "ExportListResponse",
+    "StatsResponse", "SourceStats", "SpeciesStats",
+]

backend/app/schemas/api_key.py

@@ -0,0 +1,36 @@
+from pydantic import BaseModel
+from typing import Optional
+
+
+class ApiKeyBase(BaseModel):
+    source: str
+    api_key: Optional[str] = None  # Optional for no-auth sources, used as Client ID for OAuth
+    api_secret: Optional[str] = None  # Also used as Client Secret for OAuth sources
+    access_token: Optional[str] = None  # For OAuth sources like Wikimedia
+    rate_limit_per_sec: float = 1.0
+    enabled: bool = True
+
+
+class ApiKeyCreate(ApiKeyBase):
+    pass
+
+
+class ApiKeyUpdate(BaseModel):
+    api_key: Optional[str] = None
+    api_secret: Optional[str] = None
+    access_token: Optional[str] = None
+    rate_limit_per_sec: Optional[float] = None
+    enabled: Optional[bool] = None
+
+
+class ApiKeyResponse(BaseModel):
+    id: int
+    source: str
+    api_key_masked: str  # Show only last 4 chars
+    has_secret: bool
+    has_access_token: bool
+    rate_limit_per_sec: float
+    enabled: bool
+
+    class Config:
+        from_attributes = True

backend/app/schemas/export.py

@@ -0,0 +1,45 @@
+from pydantic import BaseModel
+from datetime import datetime
+from typing import Optional, List
+
+
+class ExportFilter(BaseModel):
+    min_images_per_species: int = 100
+    licenses: Optional[List[str]] = None  # None means all
+    min_quality: Optional[float] = None
+    species_ids: Optional[List[int]] = None  # None means all
+
+
+class ExportCreate(BaseModel):
+    name: str
+    filter_criteria: ExportFilter
+    train_split: float = 0.8
+
+
+class ExportResponse(BaseModel):
+    id: int
+    name: str
+    filter_criteria: Optional[str] = None
+    train_split: float
+    status: str
+    file_path: Optional[str] = None
+    file_size: Optional[int] = None
+    species_count: Optional[int] = None
+    image_count: Optional[int] = None
+    created_at: datetime
+    completed_at: Optional[datetime] = None
+    error_message: Optional[str] = None
+
+    class Config:
+        from_attributes = True
+
+
+class ExportListResponse(BaseModel):
+    items: List[ExportResponse]
+    total: int
+
+
+class ExportPreview(BaseModel):
+    species_count: int
+    image_count: int
+    estimated_size_mb: float

backend/app/schemas/image.py

@@ -0,0 +1,47 @@
+from pydantic import BaseModel
+from datetime import datetime
+from typing import Optional, List
+
+
+class ImageBase(BaseModel):
+    species_id: int
+    source: str
+    url: str
+    license: str
+
+
+class ImageResponse(BaseModel):
+    id: int
+    species_id: int
+    species_name: Optional[str] = None
+    source: str
+    source_id: Optional[str] = None
+    url: str
+    local_path: Optional[str] = None
+    license: str
+    attribution: Optional[str] = None
+    width: Optional[int] = None
+    height: Optional[int] = None
+    quality_score: Optional[float] = None
+    status: str
+    created_at: datetime
+
+    class Config:
+        from_attributes = True
+
+
+class ImageListResponse(BaseModel):
+    items: List[ImageResponse]
+    total: int
+    page: int
+    page_size: int
+    pages: int
+
+
+class ImageFilter(BaseModel):
+    species_id: Optional[int] = None
+    source: Optional[str] = None
+    license: Optional[str] = None
+    status: Optional[str] = None
+    min_quality: Optional[float] = None
+    search: Optional[str] = None

backend/app/schemas/job.py

@@ -0,0 +1,35 @@
+from pydantic import BaseModel
+from datetime import datetime
+from typing import Optional, List
+
+
+class JobCreate(BaseModel):
+    name: str
+    source: str
+    species_ids: Optional[List[int]] = None  # None means all species
+    only_without_images: bool = False  # If True, only scrape species with 0 images
+    max_images: Optional[int] = None  # If set, only scrape species with fewer than N images
+
+
+class JobResponse(BaseModel):
+    id: int
+    name: str
+    source: str
+    species_filter: Optional[str] = None
+    status: str
+    progress_current: int
+    progress_total: int
+    images_downloaded: int
+    images_rejected: int
+    started_at: Optional[datetime] = None
+    completed_at: Optional[datetime] = None
+    error_message: Optional[str] = None
+    created_at: datetime
+
+    class Config:
+        from_attributes = True
+
+
+class JobListResponse(BaseModel):
+    items: List[JobResponse]
+    total: int

backend/app/schemas/species.py

@@ -0,0 +1,44 @@
+from pydantic import BaseModel
+from datetime import datetime
+from typing import Optional, List
+
+
+class SpeciesBase(BaseModel):
+    scientific_name: str
+    common_name: Optional[str] = None
+    genus: Optional[str] = None
+    family: Optional[str] = None
+
+
+class SpeciesCreate(SpeciesBase):
+    pass
+
+
+class SpeciesUpdate(BaseModel):
+    scientific_name: Optional[str] = None
+    common_name: Optional[str] = None
+    genus: Optional[str] = None
+    family: Optional[str] = None
+
+
+class SpeciesResponse(SpeciesBase):
+    id: int
+    created_at: datetime
+    image_count: int = 0
+
+    class Config:
+        from_attributes = True
+
+
+class SpeciesListResponse(BaseModel):
+    items: List[SpeciesResponse]
+    total: int
+    page: int
+    page_size: int
+    pages: int
+
+
+class SpeciesImportResponse(BaseModel):
+    imported: int
+    skipped: int
+    errors: List[str]

backend/app/schemas/stats.py

@@ -0,0 +1,43 @@
+from pydantic import BaseModel
+from typing import List, Dict
+
+
+class SourceStats(BaseModel):
+    source: str
+    image_count: int
+    downloaded: int
+    pending: int
+    rejected: int
+
+
+class LicenseStats(BaseModel):
+    license: str
+    count: int
+
+
+class SpeciesStats(BaseModel):
+    id: int
+    scientific_name: str
+    common_name: str | None
+    image_count: int
+
+
+class JobStats(BaseModel):
+    running: int
+    pending: int
+    completed: int
+    failed: int
+
+
+class StatsResponse(BaseModel):
+    total_species: int
+    total_images: int
+    images_downloaded: int
+    images_pending: int
+    images_rejected: int
+    disk_usage_mb: float
+    sources: List[SourceStats]
+    licenses: List[LicenseStats]
+    jobs: JobStats
+    top_species: List[SpeciesStats]
+    under_represented: List[SpeciesStats]  # Species with < 100 images