424 lines
16 KiB
Python
424 lines
16 KiB
Python
"""
|
|
FILE: app/core/chunker.py
|
|
DESCRIPTION: Zerlegt Texte in Chunks (Sliding Window oder nach Headings). Orchestriert die Smart-Edge-Allocation via SemanticAnalyzer.
|
|
VERSION: 2.9.0 (Feat: Hybrid Strict Splitting with Size Safety)
|
|
STATUS: Active
|
|
DEPENDENCIES: app.services.semantic_analyzer, app.core.derive_edges, markdown_it, yaml, asyncio
|
|
EXTERNAL_CONFIG: config/types.yaml
|
|
LAST_ANALYSIS: 2025-12-16
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
from dataclasses import dataclass
|
|
from typing import List, Dict, Optional, Tuple, Any, Set
|
|
import re
|
|
import math
|
|
import yaml
|
|
from pathlib import Path
|
|
import asyncio
|
|
import logging
|
|
|
|
# Services
|
|
from app.services.semantic_analyzer import get_semantic_analyzer
|
|
|
|
# Core Imports
|
|
try:
|
|
from app.core.derive_edges import build_edges_for_note
|
|
except ImportError:
|
|
# Mock für Tests
|
|
def build_edges_for_note(note_id, chunks, note_level_references=None, include_note_scope_refs=False): return []
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
# ==========================================
|
|
# 1. HELPER & CONFIG
|
|
# ==========================================
|
|
|
|
BASE_DIR = Path(__file__).resolve().parent.parent.parent
|
|
CONFIG_PATH = BASE_DIR / "config" / "types.yaml"
|
|
DEFAULT_PROFILE = {"strategy": "sliding_window", "target": 400, "max": 600, "overlap": (50, 80)}
|
|
_CONFIG_CACHE = None
|
|
|
|
def _load_yaml_config() -> Dict[str, Any]:
|
|
global _CONFIG_CACHE
|
|
if _CONFIG_CACHE is not None: return _CONFIG_CACHE
|
|
if not CONFIG_PATH.exists(): return {}
|
|
try:
|
|
with open(CONFIG_PATH, "r", encoding="utf-8") as f:
|
|
data = yaml.safe_load(f)
|
|
_CONFIG_CACHE = data
|
|
return data
|
|
except Exception: return {}
|
|
|
|
def get_chunk_config(note_type: str) -> Dict[str, Any]:
|
|
full_config = _load_yaml_config()
|
|
profiles = full_config.get("chunking_profiles", {})
|
|
type_def = full_config.get("types", {}).get(note_type.lower(), {})
|
|
profile_name = type_def.get("chunking_profile")
|
|
if not profile_name:
|
|
profile_name = full_config.get("defaults", {}).get("chunking_profile", "sliding_standard")
|
|
|
|
config = profiles.get(profile_name, DEFAULT_PROFILE).copy()
|
|
if "overlap" in config and isinstance(config["overlap"], list):
|
|
config["overlap"] = tuple(config["overlap"])
|
|
return config
|
|
|
|
def extract_frontmatter_from_text(md_text: str) -> Tuple[Dict[str, Any], str]:
|
|
fm_match = re.match(r'^\s*---\s*\n(.*?)\n---', md_text, re.DOTALL)
|
|
if not fm_match: return {}, md_text
|
|
try:
|
|
frontmatter = yaml.safe_load(fm_match.group(1))
|
|
if not isinstance(frontmatter, dict): frontmatter = {}
|
|
except yaml.YAMLError:
|
|
frontmatter = {}
|
|
text_without_fm = re.sub(r'^\s*---\s*\n(.*?)\n---', '', md_text, flags=re.DOTALL)
|
|
return frontmatter, text_without_fm.strip()
|
|
|
|
# ==========================================
|
|
# 2. DATA CLASSES
|
|
# ==========================================
|
|
|
|
_SENT_SPLIT = re.compile(r'(?<=[.!?])\s+(?=[A-ZÄÖÜ0-9„(])'); _WS = re.compile(r'\s+')
|
|
|
|
def estimate_tokens(text: str) -> int:
|
|
return max(1, math.ceil(len(text.strip()) / 4))
|
|
|
|
def split_sentences(text: str) -> list[str]:
|
|
text = _WS.sub(' ', text.strip())
|
|
if not text: return []
|
|
parts = _SENT_SPLIT.split(text)
|
|
return [p.strip() for p in parts if p.strip()]
|
|
|
|
@dataclass
|
|
class RawBlock:
|
|
kind: str; text: str; level: Optional[int]; section_path: str; section_title: Optional[str]
|
|
|
|
@dataclass
|
|
class Chunk:
|
|
id: str; note_id: str; index: int; text: str; window: str; token_count: int
|
|
section_title: Optional[str]; section_path: str
|
|
neighbors_prev: Optional[str]; neighbors_next: Optional[str]
|
|
suggested_edges: Optional[List[str]] = None
|
|
|
|
# ==========================================
|
|
# 3. PARSING & STRATEGIES
|
|
# ==========================================
|
|
|
|
def parse_blocks(md_text: str) -> Tuple[List[RawBlock], str]:
|
|
"""Zerlegt Text in logische Blöcke (Absätze, Header)."""
|
|
blocks = []
|
|
h1_title = "Dokument"
|
|
section_path = "/"
|
|
current_h2 = None
|
|
|
|
fm, text_without_fm = extract_frontmatter_from_text(md_text)
|
|
|
|
h1_match = re.search(r'^#\s+(.*)', text_without_fm, re.MULTILINE)
|
|
if h1_match:
|
|
h1_title = h1_match.group(1).strip()
|
|
|
|
lines = text_without_fm.split('\n')
|
|
buffer = []
|
|
|
|
for line in lines:
|
|
stripped = line.strip()
|
|
if stripped.startswith('# '):
|
|
if buffer:
|
|
content = "\n".join(buffer).strip()
|
|
if content:
|
|
blocks.append(RawBlock("paragraph", content, None, section_path, current_h2))
|
|
buffer = []
|
|
blocks.append(RawBlock("heading", stripped, 1, section_path, current_h2))
|
|
|
|
elif stripped.startswith('## '):
|
|
if buffer:
|
|
content = "\n".join(buffer).strip()
|
|
if content:
|
|
blocks.append(RawBlock("paragraph", content, None, section_path, current_h2))
|
|
buffer = []
|
|
current_h2 = stripped[3:].strip()
|
|
section_path = f"/{current_h2}"
|
|
blocks.append(RawBlock("heading", stripped, 2, section_path, current_h2))
|
|
|
|
elif stripped.startswith('### '):
|
|
if buffer:
|
|
content = "\n".join(buffer).strip()
|
|
if content:
|
|
blocks.append(RawBlock("paragraph", content, None, section_path, current_h2))
|
|
buffer = []
|
|
blocks.append(RawBlock("heading", stripped, 3, section_path, current_h2))
|
|
|
|
elif not stripped:
|
|
if buffer:
|
|
content = "\n".join(buffer).strip()
|
|
if content:
|
|
blocks.append(RawBlock("paragraph", content, None, section_path, current_h2))
|
|
buffer = []
|
|
else:
|
|
buffer.append(line)
|
|
|
|
if buffer:
|
|
content = "\n".join(buffer).strip()
|
|
if content:
|
|
blocks.append(RawBlock("paragraph", content, None, section_path, current_h2))
|
|
|
|
return blocks, h1_title
|
|
|
|
def _create_chunk_obj(chunks_list: List[Chunk], note_id: str, txt: str, win: str, sec: Optional[str], path: str):
|
|
idx = len(chunks_list)
|
|
chunks_list.append(Chunk(
|
|
id=f"{note_id}#c{idx:02d}", note_id=note_id, index=idx,
|
|
text=txt, window=win, token_count=estimate_tokens(txt),
|
|
section_title=sec, section_path=path, neighbors_prev=None, neighbors_next=None,
|
|
suggested_edges=[]
|
|
))
|
|
|
|
def _strategy_sliding_window(blocks: List[RawBlock], config: Dict[str, Any], note_id: str, doc_title: str = "", context_prefix: str = "") -> List[Chunk]:
|
|
target = config.get("target", 400)
|
|
max_tokens = config.get("max", 600)
|
|
overlap_val = config.get("overlap", (50, 80))
|
|
overlap = sum(overlap_val) // 2 if isinstance(overlap_val, tuple) else overlap_val
|
|
chunks = []; buf = []
|
|
|
|
def flush_buffer():
|
|
nonlocal buf
|
|
if not buf: return
|
|
|
|
text_body = "\n\n".join([b.text for b in buf])
|
|
win_body = f"{context_prefix}\n{text_body}".strip() if context_prefix else text_body
|
|
|
|
# Basis-Info vom ersten Block im Buffer
|
|
sec = buf[0].section_title if buf else None
|
|
path = buf[0].section_path if buf else "/"
|
|
|
|
if estimate_tokens(text_body) <= max_tokens:
|
|
_create_chunk_obj(chunks, note_id, text_body, win_body, sec, path)
|
|
else:
|
|
# Fallback: Wenn Block zu groß, intern splitten (Sentence-Level)
|
|
sentences = split_sentences(text_body)
|
|
current_chunk_sents = []
|
|
current_len = 0
|
|
|
|
for sent in sentences:
|
|
sent_len = estimate_tokens(sent)
|
|
if current_len + sent_len > target and current_chunk_sents:
|
|
c_txt = " ".join(current_chunk_sents)
|
|
c_win = f"{context_prefix}\n{c_txt}".strip() if context_prefix else c_txt
|
|
_create_chunk_obj(chunks, note_id, c_txt, c_win, sec, path)
|
|
|
|
overlap_sents = []
|
|
ov_len = 0
|
|
for s in reversed(current_chunk_sents):
|
|
if ov_len + estimate_tokens(s) < overlap:
|
|
overlap_sents.insert(0, s)
|
|
ov_len += estimate_tokens(s)
|
|
else:
|
|
break
|
|
|
|
current_chunk_sents = list(overlap_sents)
|
|
current_chunk_sents.append(sent)
|
|
current_len = ov_len + sent_len
|
|
else:
|
|
current_chunk_sents.append(sent)
|
|
current_len += sent_len
|
|
|
|
if current_chunk_sents:
|
|
c_txt = " ".join(current_chunk_sents)
|
|
c_win = f"{context_prefix}\n{c_txt}".strip() if context_prefix else c_txt
|
|
_create_chunk_obj(chunks, note_id, c_txt, c_win, sec, path)
|
|
|
|
buf = []
|
|
|
|
for b in blocks:
|
|
if b.kind == "heading":
|
|
flush_buffer()
|
|
|
|
current_buf_text = "\n\n".join([x.text for x in buf])
|
|
if buf and (estimate_tokens(current_buf_text) + estimate_tokens(b.text) >= target):
|
|
flush_buffer()
|
|
|
|
buf.append(b)
|
|
|
|
flush_buffer()
|
|
return chunks
|
|
|
|
def _strategy_by_heading(blocks: List[RawBlock], config: Dict[str, Any], note_id: str, doc_title: str = "") -> List[Chunk]:
|
|
"""
|
|
MODUS: Structured / Heading Split
|
|
- split_level: Ebene für logische Trennung (z.B. H2).
|
|
- strict_heading_split:
|
|
True: Trennt an jedem Header <= split_level.
|
|
NEU v2.9: Wenn Inhalt > max_tokens, wird trotzdem gesplittet (Safety Split).
|
|
False: Fasst zusammen bis 'target' erreicht ist.
|
|
"""
|
|
split_level = config.get("split_level", 2)
|
|
target = config.get("target", 400)
|
|
max_limit = config.get("max", 600)
|
|
strict_mode = config.get("strict_heading_split", False)
|
|
|
|
chunks = []
|
|
current_chunk_blocks = []
|
|
|
|
context_prefix = f"# {doc_title}"
|
|
|
|
def has_content(blk_list):
|
|
return any(b.kind != "heading" for b in blk_list)
|
|
|
|
def flush_current_chunk():
|
|
nonlocal current_chunk_blocks
|
|
if not current_chunk_blocks:
|
|
return
|
|
|
|
text_body = "\n\n".join([b.text for b in current_chunk_blocks])
|
|
win_body = f"{context_prefix}\n{text_body}".strip()
|
|
|
|
first_b = current_chunk_blocks[0]
|
|
sec = first_b.section_title
|
|
path = first_b.section_path
|
|
|
|
_create_chunk_obj(chunks, note_id, text_body, win_body, sec, path)
|
|
current_chunk_blocks = []
|
|
|
|
def get_current_size():
|
|
txt = "\n\n".join([b.text for b in current_chunk_blocks])
|
|
return estimate_tokens(txt)
|
|
|
|
for b in blocks:
|
|
# 1. Header Logic (Struktur-Trigger)
|
|
is_splitter = (b.kind == "heading" and b.level is not None and b.level <= split_level)
|
|
|
|
if is_splitter:
|
|
is_higher_hierarchy = (b.level < split_level)
|
|
|
|
if strict_mode:
|
|
# STRICT:
|
|
# Wir splitten immer, außer der Vor-Chunk ist leer.
|
|
if current_chunk_blocks and has_content(current_chunk_blocks):
|
|
flush_current_chunk()
|
|
current_chunk_blocks.append(b)
|
|
else:
|
|
# SOFT:
|
|
# Split bei Hierarchie-Wechsel ODER wenn voll.
|
|
if is_higher_hierarchy:
|
|
flush_current_chunk()
|
|
current_chunk_blocks.append(b)
|
|
elif current_chunk_blocks and get_current_size() >= target:
|
|
flush_current_chunk()
|
|
current_chunk_blocks.append(b)
|
|
else:
|
|
current_chunk_blocks.append(b)
|
|
else:
|
|
# 2. Content Logic (Safety Trigger für Monster-Abschnitte)
|
|
# Bevor wir den Block anhängen: Würde er das Fass zum Überlaufen bringen?
|
|
# Wir nutzen hier 'max' als harte Grenze für den Safety-Split.
|
|
current_size = get_current_size()
|
|
block_size = estimate_tokens(b.text)
|
|
|
|
if current_chunk_blocks and (current_size + block_size > max_limit):
|
|
# NOTBREMSE: Chunk wird zu groß.
|
|
# Wir splitten hier, auch wenn kein Header da ist.
|
|
# Der Kontext (Section Title) bleibt erhalten, da er aus `current_h2` kommt (siehe parse_blocks).
|
|
flush_current_chunk()
|
|
current_chunk_blocks.append(b)
|
|
else:
|
|
current_chunk_blocks.append(b)
|
|
|
|
# Letzten Rest flushen
|
|
flush_current_chunk()
|
|
|
|
return chunks
|
|
|
|
# ==========================================
|
|
# 4. ORCHESTRATION (ASYNC)
|
|
# ==========================================
|
|
|
|
async def assemble_chunks(note_id: str, md_text: str, note_type: str, config: Optional[Dict] = None) -> List[Chunk]:
|
|
if config is None:
|
|
config = get_chunk_config(note_type)
|
|
|
|
fm, body_text = extract_frontmatter_from_text(md_text)
|
|
note_status = fm.get("status", "").lower()
|
|
|
|
primary_strategy = config.get("strategy", "sliding_window")
|
|
enable_smart_edges = config.get("enable_smart_edge_allocation", False)
|
|
|
|
if enable_smart_edges and note_status in ["draft", "initial_gen"]:
|
|
logger.info(f"Chunker: Skipping Smart Edges for draft '{note_id}'.")
|
|
enable_smart_edges = False
|
|
|
|
blocks, doc_title = parse_blocks(md_text)
|
|
|
|
if primary_strategy == "by_heading":
|
|
chunks = await asyncio.to_thread(_strategy_by_heading, blocks, config, note_id, doc_title)
|
|
else:
|
|
chunks = await asyncio.to_thread(_strategy_sliding_window, blocks, config, note_id, doc_title)
|
|
|
|
if not chunks:
|
|
return []
|
|
|
|
if enable_smart_edges:
|
|
chunks = await _run_smart_edge_allocation(chunks, md_text, note_id, note_type)
|
|
|
|
for i, ch in enumerate(chunks):
|
|
ch.neighbors_prev = chunks[i-1].id if i > 0 else None
|
|
ch.neighbors_next = chunks[i+1].id if i < len(chunks)-1 else None
|
|
|
|
return chunks
|
|
|
|
def _extract_all_edges_from_md(md_text: str, note_id: str, note_type: str) -> List[str]:
|
|
dummy_chunk = {
|
|
"chunk_id": f"{note_id}#full",
|
|
"text": md_text,
|
|
"content": md_text,
|
|
"window": md_text,
|
|
"type": note_type
|
|
}
|
|
raw_edges = build_edges_for_note(
|
|
note_id,
|
|
[dummy_chunk],
|
|
note_level_references=None,
|
|
include_note_scope_refs=False
|
|
)
|
|
all_candidates = set()
|
|
for e in raw_edges:
|
|
kind = e.get("kind")
|
|
target = e.get("target_id")
|
|
if target and kind not in ["belongs_to", "next", "prev", "backlink"]:
|
|
all_candidates.add(f"{kind}:{target}")
|
|
return list(all_candidates)
|
|
|
|
async def _run_smart_edge_allocation(chunks: List[Chunk], full_text: str, note_id: str, note_type: str) -> List[Chunk]:
|
|
analyzer = get_semantic_analyzer()
|
|
candidate_list = _extract_all_edges_from_md(full_text, note_id, note_type)
|
|
|
|
if not candidate_list:
|
|
return chunks
|
|
|
|
tasks = []
|
|
for chunk in chunks:
|
|
tasks.append(analyzer.assign_edges_to_chunk(chunk.text, candidate_list, note_type))
|
|
|
|
results_per_chunk = await asyncio.gather(*tasks)
|
|
|
|
assigned_edges_global = set()
|
|
|
|
for i, confirmed_edges in enumerate(results_per_chunk):
|
|
chunk = chunks[i]
|
|
chunk.suggested_edges = confirmed_edges
|
|
assigned_edges_global.update(confirmed_edges)
|
|
|
|
if confirmed_edges:
|
|
injection_str = "\n" + " ".join([f"[[rel:{e.split(':')[0]}|{e.split(':')[1]}]]" for e in confirmed_edges if ':' in e])
|
|
chunk.text += injection_str
|
|
chunk.window += injection_str
|
|
|
|
unassigned = set(candidate_list) - assigned_edges_global
|
|
if unassigned:
|
|
fallback_str = "\n" + " ".join([f"[[rel:{e.split(':')[0]}|{e.split(':')[1]}]]" for e in unassigned if ':' in e])
|
|
for chunk in chunks:
|
|
chunk.text += fallback_str
|
|
chunk.window += fallback_str
|
|
if chunk.suggested_edges is None: chunk.suggested_edges = []
|
|
chunk.suggested_edges.extend(list(unassigned))
|
|
|
|
return chunks |