fix(aym,uyusmazlik): adapt to rebuilt AYM and Uyuşmazlık sites

Both sites were rebuilt and their old endpoints now 404:
- AYM moved to a single-page app backed by a JSON API
  (POST /api/core/public/search, kararTipi NormDenetimi/BireyselBasvuru;
  full text via {id, size:1} -> "icerik" HTML). Old /Ara, /ND/, /BB/ gone.
- Uyuşmazlık moved to ASP.NET WebForms (viewstate postback to /, GridView
  results, decisions served as /Uploads/{EsasNo}.pdf). Old /Arama/Search gone.

Changes:
- New anayasa_mcp_module/api_client.py: shared KBB JSON client, base64url
  document-id codec, HTML->Markdown + HTML text stripping helpers.
- Rewrite anayasa client/bireysel_client/unified_client over the new API,
  preserving public method names and response models.
- Rewrite uyusmazlik client for the postback flow + GridView parse + PDF
  document conversion; simplify request model to text + scope + paging.
- Trim search_anayasa_unified and search_uyusmazlik_decisions tool signatures
  to parameters the new APIs actually support; drop dead enums.

Verified live via FastMCP client: search + document retrieval work for AYM
norm/bireysel and Uyuşmazlık.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
saidsurucu
2026-06-30 19:22:49 +03:00
co-authored by Claude Opus 4.8
parent 6bbc656dc6
commit fadc3b0bc0
9 changed files with 634 additions and 1116 deletions
+125 -197
View File
@@ -1,251 +1,179 @@
# uyusmazlik_mcp_module/client.py
#
# Client for the rebuilt Uyuşmazlık Mahkemesi search site
# (https://kararlar.uyusmazlik.gov.tr). The site is an ASP.NET WebForms app:
# searching is a form postback against "/" that returns an HTML page with a
# GridView of results, and each decision is a PDF served from /Uploads/.
#
# The previous AJAX endpoint (/Arama/Search) was retired and now returns 404.
import asyncio
import io
import logging
import re
from typing import Dict, List, Optional
from urllib.parse import urljoin
import httpx
from bs4 import BeautifulSoup
from typing import Dict, Any, List, Optional, Union, Tuple
import logging
import html
import re
import io
from markitdown import MarkItDown
from urllib.parse import urljoin
from .models import (
UyusmazlikSearchRequest,
UyusmazlikApiDecisionEntry,
UyusmazlikSearchResponse,
UyusmazlikDocumentMarkdown,
UyusmazlikBolumEnum,
UyusmazlikTuruEnum,
UyusmazlikKararSonucuEnum
)
logger = logging.getLogger(__name__)
if not logger.hasHandlers():
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(name)s - %(levelname)s - %(message)s')
# --- Mappings from user-friendly Enum values to API IDs ---
BOLUM_ENUM_TO_ID_MAP = {
UyusmazlikBolumEnum.CEZA_BOLUMU: "f6b74320-f2d7-4209-ad6e-c6df180d4e7c",
UyusmazlikBolumEnum.GENEL_KURUL_KARARLARI: "e4ca658d-a75a-4719-b866-b2d2f1c3b1d9",
UyusmazlikBolumEnum.HUKUK_BOLUMU: "96b26fc4-ef8e-4a4f-a9cc-a3de89952aa1",
UyusmazlikBolumEnum.TUMU: "", # Represents "...Seçiniz..." or all - empty string for API
"ALL": "" # Also map the new "ALL" literal to empty string for backward compatibility
}
# ASP.NET hidden fields that must be round-tripped on every postback.
_HIDDEN_FIELDS = ("__VIEWSTATE", "__VIEWSTATEGENERATOR", "__EVENTVALIDATION")
UYUSMAZLIK_TURU_ENUM_TO_ID_MAP = {
UyusmazlikTuruEnum.GOREV_UYUSMAZLIGI: "7b1e2cd3-8f09-418a-921c-bbe501e1740c",
UyusmazlikTuruEnum.HUKUM_UYUSMAZLIGI: "19b88402-172b-4c1d-8339-595c942a89f5",
UyusmazlikTuruEnum.TUMU: "", # Represents "...Seçiniz..." or all - empty string for API
"ALL": "" # Also map the new "ALL" literal to empty string for backward compatibility
}
KARAR_SONUCU_ENUM_TO_ID_MAP = {
# These IDs are from the form HTML provided by the user
UyusmazlikKararSonucuEnum.HUKUM_UYUSMAZLIGI_OLMADIGINA_DAIR: "6f47d87f-dcb5-412e-9878-000385dba1d9",
UyusmazlikKararSonucuEnum.HUKUM_UYUSMAZLIGI_OLDUGUNA_DAIR: "5a01742a-c440-4c4a-ba1f-da20837cffed",
# Add all other 'Karar Sonucu' enum members and their corresponding GUIDs
# by inspecting the 'KararSonucuList' checkboxes in the provided form HTML.
}
# --- End Mappings ---
class UyusmazlikApiClient:
BASE_URL = "https://kararlar.uyusmazlik.gov.tr"
SEARCH_ENDPOINT = "/Arama/Search"
# Individual documents are fetched by their full URLs obtained from search results.
SEARCH_PATH = "/"
def __init__(self, request_timeout: float = 30.0):
self.request_timeout = request_timeout
# Create shared httpx client for all requests
# A persistent cookie-aware client so ASP.NET session/viewstate are kept.
self.http_client = httpx.AsyncClient(
base_url=self.BASE_URL,
headers={
"Accept": "*/*",
"Accept-Encoding": "gzip, deflate, br, zstd",
"Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
"Accept-Language": "tr-TR,tr;q=0.9,en-US;q=0.8,en;q=0.7",
"X-Requested-With": "XMLHttpRequest",
"User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) "
"AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36",
"Origin": self.BASE_URL,
"Referer": self.BASE_URL + "/",
},
timeout=request_timeout,
verify=False
verify=False,
follow_redirects=True,
)
@staticmethod
def _extract_hidden_fields(html_content: str) -> Dict[str, str]:
soup = BeautifulSoup(html_content, "html.parser")
fields: Dict[str, str] = {}
for name in _HIDDEN_FIELDS:
tag = soup.find("input", attrs={"name": name})
fields[name] = tag["value"] if tag and tag.has_attr("value") else ""
return fields
async def search_decisions(
self,
params: UyusmazlikSearchRequest
) -> UyusmazlikSearchResponse:
bolum_id_for_api = BOLUM_ENUM_TO_ID_MAP.get(params.bolum, "")
uyusmazlik_id_for_api = UYUSMAZLIK_TURU_ENUM_TO_ID_MAP.get(params.uyusmazlik_turu, "")
form_data_list: List[Tuple[str, str]] = []
@staticmethod
def _parse_results(html_content: str, base_url: str) -> UyusmazlikSearchResponse:
soup = BeautifulSoup(html_content, "html.parser")
def add_to_form_data(key: str, value: Optional[str]):
# API expects empty strings for omitted optional fields based on user payload example
form_data_list.append((key, value or ""))
decisions: List[UyusmazlikApiDecisionEntry] = []
grid = soup.find("table", id="GridView1")
if grid:
rows = grid.find_all("tr")
for row in rows[1:]: # skip header row
cells = row.find_all("td")
if len(cells) < 4:
continue
# The İşlemler cell holds the PDF "Görüntüle" link. Pager rows also
# contain <a> tags (javascript:__doPostBack ...), so require a real
# document link and skip everything else.
link_tag = cells[3].find(
"a", href=lambda h: h and not h.strip().lower().startswith("javascript:")
)
if not link_tag:
continue
href = link_tag["href"].strip()
if "uploads" not in href.lower() and not href.lower().endswith(".pdf"):
continue
document_url = urljoin(base_url + "/", href)
decisions.append(UyusmazlikApiDecisionEntry(
esas_sayisi=cells[0].get_text(strip=True) or None,
karar_sayisi=cells[1].get_text(strip=True) or None,
karar_tarihi=cells[2].get_text(strip=True) or None,
document_url=document_url,
))
add_to_form_data("BolumId", bolum_id_for_api)
add_to_form_data("UyusmazlikId", uyusmazlik_id_for_api)
if params.karar_sonuclari:
for enum_member in params.karar_sonuclari:
api_id = KARAR_SONUCU_ENUM_TO_ID_MAP.get(enum_member)
if api_id: # Only add if a valid ID is found
form_data_list.append(('KararSonucuList', api_id))
add_to_form_data("EsasYil", params.esas_yil)
add_to_form_data("EsasSayisi", params.esas_sayisi)
add_to_form_data("KararYil", params.karar_yil)
add_to_form_data("KararSayisi", params.karar_sayisi)
add_to_form_data("KanunNo", params.kanun_no)
add_to_form_data("KararDateBegin", params.karar_date_begin)
add_to_form_data("KararDateEnd", params.karar_date_end)
add_to_form_data("ResmiGazeteSayi", params.resmi_gazete_sayi)
add_to_form_data("ResmiGazeteDate", params.resmi_gazete_date)
add_to_form_data("Icerik", params.icerik)
add_to_form_data("Tumce", params.tumce)
add_to_form_data("WildCard", params.wild_card)
add_to_form_data("Hepsi", params.hepsi)
add_to_form_data("Herhangibirisi", params.herhangi_birisi)
add_to_form_data("NotHepsi", params.not_hepsi)
# Try to read a "N kayıt/sonuç/karar bulundu" style count if present.
total_records: Optional[int] = None
count_match = re.search(r'(\d+)\s*(?:adet\s*)?(?:kayıt|sonuç|karar)\b', html_content, re.IGNORECASE)
if count_match:
total_records = int(count_match.group(1))
# Convert form data to dict for httpx
form_data_dict = {}
for key, value in form_data_list:
if key in form_data_dict:
# Handle multiple values (like KararSonucuList)
if not isinstance(form_data_dict[key], list):
form_data_dict[key] = [form_data_dict[key]]
form_data_dict[key].append(value)
else:
form_data_dict[key] = value
return UyusmazlikSearchResponse(decisions=decisions, total_records_found=total_records)
logger.info(f"UyusmazlikApiClient (httpx): Performing search to {self.SEARCH_ENDPOINT} with form_data: {form_data_dict}")
try:
# Use shared httpx client
response = await self.http_client.post(
self.SEARCH_ENDPOINT,
data=form_data_dict,
headers={"Content-Type": "application/x-www-form-urlencoded; charset=UTF-8"}
async def search_decisions(self, params: UyusmazlikSearchRequest) -> UyusmazlikSearchResponse:
# 1. Load the landing page to obtain a fresh viewstate + session cookie.
landing = await self.http_client.get(self.SEARCH_PATH)
landing.raise_for_status()
form_data = self._extract_hidden_fields(landing.text)
# 2. Submit the search form.
form_data.update({
"txtSearch": params.icerik or "",
"rblSearchScope": params.search_scope,
"btnSearch": "Ara",
})
if params.case_sensitive:
form_data["chkCaseSensitive"] = "on"
logger.info("UyusmazlikApiClient: search icerik=%r scope=%s page=%s",
params.icerik, params.search_scope, params.page_number)
response = await self.http_client.post(
self.SEARCH_PATH,
data=form_data,
headers={"Content-Type": "application/x-www-form-urlencoded"},
)
response.raise_for_status()
html_content = response.text
# 3. Navigate the GridView pager if a later page is requested.
if params.page_number > 1:
page_fields = self._extract_hidden_fields(html_content)
page_fields.update({
"txtSearch": params.icerik or "",
"rblSearchScope": params.search_scope,
"__EVENTTARGET": "GridView1",
"__EVENTARGUMENT": f"Page${params.page_number}",
})
if params.case_sensitive:
page_fields["chkCaseSensitive"] = "on"
page_response = await self.http_client.post(
self.SEARCH_PATH,
data=page_fields,
headers={"Content-Type": "application/x-www-form-urlencoded"},
)
response.raise_for_status()
html_content = response.text
logger.debug("UyusmazlikApiClient (httpx): Received HTML response for search.")
except httpx.HTTPError as e:
logger.error(f"UyusmazlikApiClient (httpx): HTTP client error during search: {e}")
raise # Re-raise to be handled by the MCP tool
except Exception as e:
logger.error(f"UyusmazlikApiClient (httpx): Error processing search request: {e}")
raise
page_response.raise_for_status()
html_content = page_response.text
# --- HTML Parsing (remains the same as previous version) ---
soup = BeautifulSoup(html_content, 'html.parser')
total_records_text_div = soup.find("div", class_="pull-right label label-important")
total_records = None
if total_records_text_div:
match_records = re.search(r'(\d+)\s*adet kayıt bulundu', total_records_text_div.get_text(strip=True))
if match_records:
total_records = int(match_records.group(1))
result_table = soup.find("table", class_="table-hover")
processed_decisions: List[UyusmazlikApiDecisionEntry] = []
if result_table:
rows = result_table.find_all("tr")
if len(rows) > 1: # Skip header row
for row in rows[1:]:
cols = row.find_all('td')
if len(cols) >= 5:
try:
popover_div = cols[0].find("div", attrs={"data-rel": "popover"})
popover_content_raw = popover_div["data-content"] if popover_div and popover_div.has_attr("data-content") else None
link_tag = cols[0].find('a')
doc_relative_url = link_tag['href'] if link_tag and link_tag.has_attr('href') else None
if not doc_relative_url: continue
document_url_str = urljoin(self.BASE_URL, doc_relative_url)
return self._parse_results(html_content, self.BASE_URL)
pdf_link_tag = cols[5].find('a', href=re.compile(r'\.pdf$', re.IGNORECASE)) if len(cols) > 5 else None
pdf_url_str = urljoin(self.BASE_URL, pdf_link_tag['href']) if pdf_link_tag and pdf_link_tag.has_attr('href') else None
decision_data_parsed = {
"karar_sayisi": cols[0].get_text(strip=True),
"esas_sayisi": cols[1].get_text(strip=True),
"bolum": cols[2].get_text(strip=True),
"uyusmazlik_konusu": cols[3].get_text(strip=True),
"karar_sonucu": cols[4].get_text(strip=True),
"popover_content": html.unescape(popover_content_raw) if popover_content_raw else None,
"document_url": document_url_str,
"pdf_url": pdf_url_str
}
decision_model = UyusmazlikApiDecisionEntry(**decision_data_parsed)
processed_decisions.append(decision_model)
except Exception as e:
logger.warning(f"UyusmazlikApiClient: Could not parse decision row. Row content: {row.get_text(strip=True, separator=' | ')}, Error: {e}")
return UyusmazlikSearchResponse(
decisions=processed_decisions,
total_records_found=total_records
)
def _convert_html_to_markdown_uyusmazlik(self, full_decision_html_content: str) -> Optional[str]:
"""Converts direct HTML content (from an Uyuşmazlık decision page) to Markdown."""
if not full_decision_html_content:
return None
processed_html = html.unescape(full_decision_html_content)
# As per user request, pass the full (unescaped) HTML to MarkItDown
html_input_for_markdown = processed_html
markdown_text = None
def _convert_pdf_to_markdown(self, pdf_bytes: bytes) -> Optional[str]:
try:
# Convert HTML string to bytes and create BytesIO stream
html_bytes = html_input_for_markdown.encode('utf-8')
html_stream = io.BytesIO(html_bytes)
# Pass BytesIO stream to MarkItDown to avoid temp file creation
md_converter = MarkItDown()
conversion_result = md_converter.convert(html_stream)
markdown_text = conversion_result.text_content
logger.info("UyusmazlikApiClient: HTML to Markdown conversion successful.")
pdf_stream = io.BytesIO(pdf_bytes)
conversion_result = MarkItDown().convert(pdf_stream, file_extension=".pdf")
return conversion_result.text_content
except Exception as e:
logger.error(f"UyusmazlikApiClient: Error during MarkItDown HTML to Markdown conversion: {e}")
return markdown_text
logger.error("UyusmazlikApiClient: PDF to Markdown conversion error: %s", e)
return None
async def get_decision_document_as_markdown(self, document_url: str) -> UyusmazlikDocumentMarkdown:
"""
Retrieves a specific Uyuşmazlık decision from its full URL and returns content as Markdown.
"""
logger.info(f"UyusmazlikApiClient (httpx for docs): Fetching Uyuşmazlık document for Markdown from URL: {document_url}")
"""Fetch an Uyuşmazlık decision PDF and return its content as Markdown."""
logger.info("UyusmazlikApiClient: Fetching document PDF from %s", document_url)
try:
# Using a new httpx.AsyncClient instance for this GET request for simplicity
async with httpx.AsyncClient(verify=False, timeout=self.request_timeout) as doc_fetch_client:
get_response = await doc_fetch_client.get(document_url, headers={"Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8"})
get_response.raise_for_status()
html_content_from_api = get_response.text
if not isinstance(html_content_from_api, str) or not html_content_from_api.strip():
logger.warning(f"UyusmazlikApiClient: Received empty or non-string HTML from URL {document_url}.")
return UyusmazlikDocumentMarkdown(source_url=document_url, markdown_content=None)
markdown_content = await asyncio.to_thread(self._convert_html_to_markdown_uyusmazlik, html_content_from_api)
response = await self.http_client.get(
document_url,
headers={"Accept": "application/pdf,*/*"},
)
response.raise_for_status()
markdown_content = await asyncio.to_thread(self._convert_pdf_to_markdown, response.content)
return UyusmazlikDocumentMarkdown(source_url=document_url, markdown_content=markdown_content)
except httpx.RequestError as e:
logger.error(f"UyusmazlikApiClient (httpx for docs): HTTP error fetching Uyuşmazlık document from {document_url}: {e}")
raise
except Exception as e:
logger.error(f"UyusmazlikApiClient (httpx for docs): General error processing Uyuşmazlık document from {document_url}: {e}")
except httpx.HTTPError as e:
logger.error("UyusmazlikApiClient: HTTP error fetching document from %s: %s", document_url, e)
raise
async def close_client_session(self):
"""Close the shared httpx client session."""
if hasattr(self, 'http_client') and self.http_client:
if hasattr(self, "http_client") and self.http_client and not self.http_client.is_closed:
await self.http_client.aclose()
logger.info("UyusmazlikApiClient: HTTP client session closed.")
else:
logger.info("UyusmazlikApiClient: No persistent client session from __init__ to close.")
+27 -72
View File
@@ -1,86 +1,41 @@
# uyusmazlik_mcp_module/models.py
from pydantic import BaseModel, Field, HttpUrl
from typing import List, Optional
from enum import Enum
from typing import List, Optional, Literal
# Enum definitions for user-friendly input based on the provided HTML form
class UyusmazlikBolumEnum(str, Enum):
"""User-friendly names for 'BolumId'."""
TUMU = "ALL" # Represents "...Seçiniz..." or all
CEZA_BOLUMU = "Ceza Bölümü"
GENEL_KURUL_KARARLARI = "Genel Kurul Kararları"
HUKUK_BOLUMU = "Hukuk Bölümü"
# The Uyuşmazlık Mahkemesi search site was rebuilt as an ASP.NET WebForms app.
# It now offers only a single free-text search with a scope selector; the old
# Bölüm / Uyuşmazlık Türü / Karar Sonucu / Esas-Karar year filters no longer exist.
class UyusmazlikTuruEnum(str, Enum):
"""User-friendly names for 'UyusmazlikId'."""
TUMU = "ALL" # Represents "...Seçiniz..." or all
GOREV_UYUSMAZLIGI = "Görev Uyuşmazlığı"
HUKUM_UYUSMAZLIGI = "Hüküm Uyuşmazlığı"
UyusmazlikSearchScope = Literal["All", "EsasNo", "KararNo"]
class UyusmazlikKararSonucuEnum(str, Enum): # Based on checkbox text in the form
"""User-friendly names for 'KararSonucuList' items."""
HUKUM_UYUSMAZLIGI_OLMADIGINA_DAIR = "Hüküm Uyuşmazlığı Olmadığına Dair"
HUKUM_UYUSMAZLIGI_OLDUGUNA_DAIR = "Hüküm Uyuşmazlığı Olduğuna Dair"
# Add other "Karar Sonucu" options from the form's checkboxes as Enum members
# Example: GOREVLI_YARGI_YERI_ADLI = "Görevli Yargı Yeri Belirlenmesine Dair (Adli Yargı)"
# The client will map these enum values (which are strings) to their respective IDs.
class UyusmazlikSearchRequest(BaseModel): # This is the model the MCP tool will accept
"""Model for Uyuşmazlık Mahkemesi search request using user-friendly terms."""
icerik: Optional[str] = Field("", description="Search text")
bolum: Optional[UyusmazlikBolumEnum] = Field(
UyusmazlikBolumEnum.TUMU,
description="Department"
class UyusmazlikSearchRequest(BaseModel):
"""Model for the Uyuşmazlık Mahkemesi search request."""
icerik: str = Field("", description="Search text (txtSearch).")
search_scope: UyusmazlikSearchScope = Field(
"All",
description="Search scope: 'All' (full text), 'EsasNo' (by case number), 'KararNo' (by decision number).",
)
uyusmazlik_turu: Optional[UyusmazlikTuruEnum] = Field(
UyusmazlikTuruEnum.TUMU,
description="Dispute type"
)
# User provides a list of user-friendly names for Karar Sonucu
karar_sonuclari: Optional[List[UyusmazlikKararSonucuEnum]] = Field( # Changed to list of Enums
default_factory=list,
description="Decision types"
)
esas_yil: Optional[str] = Field("", description="Case year")
esas_sayisi: Optional[str] = Field("", description="Case no")
karar_yil: Optional[str] = Field("", description="Decision year")
karar_sayisi: Optional[str] = Field("", description="Decision no")
kanun_no: Optional[str] = Field("", description="Law no")
karar_date_begin: Optional[str] = Field("", description="Start date (DD.MM.YYYY)")
karar_date_end: Optional[str] = Field("", description="End date (DD.MM.YYYY)")
resmi_gazete_sayi: Optional[str] = Field("", description="Gazette no")
resmi_gazete_date: Optional[str] = Field("", description="Gazette date (DD.MM.YYYY)")
# Detailed text search fields from the "icerikDetail" section of the form
tumce: Optional[str] = Field("", description="Exact phrase")
wild_card: Optional[str] = Field("", description="Wildcard search")
hepsi: Optional[str] = Field("", description="All words")
herhangi_birisi: Optional[str] = Field("", description="Any word")
not_hepsi: Optional[str] = Field("", description="Exclude words")
case_sensitive: bool = Field(False, description="Whether the search is case sensitive (chkCaseSensitive).")
page_number: int = Field(1, ge=1, description="Result page number (GridView pager).")
class UyusmazlikApiDecisionEntry(BaseModel):
"""Model for an individual decision entry parsed from Uyuşmazlık API's HTML search response."""
karar_sayisi: Optional[str] = Field(None)
esas_sayisi: Optional[str] = Field(None)
bolum: Optional[str] = Field(None)
uyusmazlik_konusu: Optional[str] = Field(None)
karar_sonucu: Optional[str] = Field(None)
popover_content: Optional[str] = Field(None, description="Summary")
document_url: HttpUrl # Full URL to the decision document HTML page
pdf_url: Optional[HttpUrl] = Field(None, description="PDF URL")
"""A single decision row parsed from the Uyuşmazlık GridView results."""
esas_sayisi: Optional[str] = Field(None, description="Case number (Esas No).")
karar_sayisi: Optional[str] = Field(None, description="Decision number (Karar No).")
karar_tarihi: Optional[str] = Field(None, description="Decision date (DD/MM/YYYY).")
document_url: HttpUrl = Field(..., description="Full URL to the decision PDF document.")
class UyusmazlikSearchResponse(BaseModel): # This is what the MCP tool will return
"""Response model for Uyuşmazlık Mahkemesi search results for the MCP tool."""
class UyusmazlikSearchResponse(BaseModel):
"""Response model for Uyuşmazlık Mahkemesi search results."""
decisions: List[UyusmazlikApiDecisionEntry]
total_records_found: Optional[int] = Field(None, description="Total number of records found for the query, if available.")
total_records_found: Optional[int] = Field(None, description="Total number of records found, if reported.")
class UyusmazlikDocumentMarkdown(BaseModel):
"""Model for an Uyuşmazlık decision document, containing only Markdown content."""
source_url: HttpUrl # The URL from which the content was fetched
markdown_content: Optional[str] = Field(None, description="The decision content converted to Markdown.")
"""Model for an Uyuşmazlık decision document, containing Markdown content."""
source_url: HttpUrl
markdown_content: Optional[str] = Field(None, description="The decision PDF content converted to Markdown.")