update
This commit is contained in:
@@ -24,11 +24,11 @@ from jsonio import read_json_file as _json_datei
|
||||
from paths import arbeit_dir, bausteine_path, project_dir, subbausteine_path, quelle_path, quelle_crawl_dir, safe_ordner
|
||||
from crawl import crawl
|
||||
from pipeline import (
|
||||
CANCELLED, FAILED, GenContext, _extra, _log, _prompt, _race, _rest_schema,
|
||||
CANCELLED, FAILED, GenContext, _extra, _log, _prompt, _race, _relevanz_schema, _rest_schema,
|
||||
_runde_schema, _semaphore, _str_liste, _stufen_schema, _timeout, run_single_slot,
|
||||
)
|
||||
from textkit import (
|
||||
_eindeutige_titel, _norm_titel, _parse_auswahl, _parse_subbausteine, _titel,
|
||||
_eindeutige_titel, _lade_bausteine, _norm_titel, _parse_auswahl, _parse_subbausteine, _titel,
|
||||
_titel_aufloesen, _titel_index, _vormerge,
|
||||
)
|
||||
|
||||
@@ -50,6 +50,7 @@ BAUSTEINE_STEPS = (
|
||||
"Recherche", "Konsolidierung", "Klärung",
|
||||
"Subbausteine finden", "Subbausteine wählen", "Subbausteine klären",
|
||||
"Stufen finden", "Stufen wählen", "Stufen klären",
|
||||
"Relevanz finden", "Relevanz wählen", "Relevanz klären",
|
||||
)
|
||||
|
||||
|
||||
@@ -78,6 +79,36 @@ def _crawl_fertig(topic: str) -> bool:
|
||||
return (quelle_crawl_dir(topic) / ".done").exists() # Marker erst bei sauberem Abschluss
|
||||
|
||||
|
||||
_STUFEN = ("einfach", "mittel", "schwer")
|
||||
|
||||
|
||||
def lade_uebersicht(topic: str) -> list[dict]:
|
||||
"""Strukturierte Baustein-Liste für die Übersicht: Titel + Beschreibung + Subbausteine/Stufen.
|
||||
|
||||
Verbindet bausteine.md (Nummer/Titel/Beschreibung) mit der Sidecar subbausteine.json
|
||||
(Key = Titel). Fehlt die Sidecar, sind die Subbaustein-Listen leer.
|
||||
"""
|
||||
entries = _lade_bausteine(_read(bausteine_path(topic)))
|
||||
sidecar = _json_datei(subbausteine_path(topic))
|
||||
sidecar = sidecar if isinstance(sidecar, dict) else {}
|
||||
out = []
|
||||
for num, entry in entries.items():
|
||||
titel = _titel(entry)
|
||||
teile = entry.split(" — ", 1)
|
||||
beschreibung = teile[1].strip() if len(teile) == 2 else ""
|
||||
subbausteine = [
|
||||
{
|
||||
"titel": t,
|
||||
"stufe": s.get("stufe") if s.get("stufe") in _STUFEN else "mittel",
|
||||
"relevanz": s.get("relevanz") if s.get("relevanz") in ("relevant", "rand") else None,
|
||||
}
|
||||
for s in (sidecar.get(titel) or [])
|
||||
if isinstance(s, dict) and (t := str(s.get("titel", "")).strip())
|
||||
]
|
||||
out.append({"num": num, "titel": titel, "beschreibung": beschreibung, "subbausteine": subbausteine})
|
||||
return out
|
||||
|
||||
|
||||
def _bausteine_steps(topic: str) -> tuple:
|
||||
"""Schritte je Quelle: link bekommt vorne „Quelle laden", projekt zusätzlich „Ergänzung".
|
||||
|
||||
@@ -89,6 +120,7 @@ def _bausteine_steps(topic: str) -> tuple:
|
||||
rest = (
|
||||
"Subbausteine finden", "Subbausteine wählen", "Subbausteine klären",
|
||||
"Stufen finden", "Stufen wählen", "Stufen klären",
|
||||
"Relevanz finden", "Relevanz wählen", "Relevanz klären",
|
||||
)
|
||||
mitte = base + (("Ergänzung",) if q["type"] == "projekt" else ()) + rest
|
||||
return (("Quelle laden",) if q["type"] == "link" else ()) + mitte
|
||||
@@ -117,7 +149,7 @@ def _bausteine_files(topic: str) -> dict:
|
||||
def _alle_slot_dateien(files: dict) -> list[Path]:
|
||||
arbeit = files["arbeit"]
|
||||
# Subbaustein-/Stufen-Slots sind pro Chunk dynamisch — per Glob einsammeln.
|
||||
dyn = (list(arbeit.glob("subbaustein-*")) + list(arbeit.glob("stufe-*"))) if arbeit.is_dir() else []
|
||||
dyn = (list(arbeit.glob("subbaustein-*")) + list(arbeit.glob("stufe-*")) + list(arbeit.glob("relevanz-*"))) if arbeit.is_dir() else []
|
||||
return [
|
||||
*files["recherche"], files["recherche_mapping"],
|
||||
*(p for slots in files["auswahl"].values() for p in slots),
|
||||
@@ -153,8 +185,12 @@ def _resume_step(topic: str) -> int:
|
||||
return _step_idx(topic, "Klärung")
|
||||
if q["type"] == "projekt" and not files["ergaenzung"].exists():
|
||||
return _step_idx(topic, "Ergänzung")
|
||||
if _sidecar_schema(_json_datei(files["sidecar"])) is not None:
|
||||
return len(_bausteine_steps(topic))
|
||||
sidecar = _json_datei(files["sidecar"])
|
||||
if _sidecar_schema(sidecar) is not None:
|
||||
# Stufen fertig; nur noch Relevanz offen?
|
||||
if _relevanz_komplett(sidecar):
|
||||
return len(_bausteine_steps(topic))
|
||||
return _step_idx(topic, "Relevanz finden")
|
||||
if _sub_roh_schema(_json_datei(files["sub_roh"])) is None:
|
||||
return _step_idx(topic, "Subbausteine finden")
|
||||
return _step_idx(topic, "Stufen finden")
|
||||
@@ -287,7 +323,7 @@ def _sub_roh_schema(data):
|
||||
|
||||
|
||||
def _sidecar_schema(data):
|
||||
"""{Baustein-Titel: [{titel, stufe}, …]} → dict · sonst None (finale Sidecar)."""
|
||||
"""{Baustein-Titel: [{titel, stufe}, …]} → dict · sonst None (Sidecar mit Stufen)."""
|
||||
if not isinstance(data, dict) or not data:
|
||||
return None
|
||||
for v in data.values():
|
||||
@@ -299,6 +335,17 @@ def _sidecar_schema(data):
|
||||
return data
|
||||
|
||||
|
||||
def _relevanz_komplett(data) -> bool:
|
||||
"""Jeder Subbaustein der Sidecar trägt eine gültige Relevanz (relevant/rand)?"""
|
||||
if not isinstance(data, dict) or not data:
|
||||
return False
|
||||
return all(
|
||||
isinstance(s, dict) and s.get("relevanz") in ("relevant", "rand")
|
||||
for v in data.values() if isinstance(v, list)
|
||||
for s in v
|
||||
)
|
||||
|
||||
|
||||
def _read(p: Path) -> str:
|
||||
return p.read_text(encoding="utf-8") if p.exists() else ""
|
||||
|
||||
@@ -574,6 +621,115 @@ async def _stufen_block(ctx: GenContext, set_p, files: dict, roh: dict, instruct
|
||||
return sidecar
|
||||
|
||||
|
||||
async def _relevanz_block(ctx: GenContext, set_p, files: dict, sidecar: dict, instructions: str) -> dict | None:
|
||||
"""Block D: drei Phasen mit Barriere — Finden (relevant/rand), Wählen (Vote), Klären.
|
||||
Items aus der Sidecar; lokale IDs 1..n pro Paket → globale gid.
|
||||
→ {gid: relevanz} oder None bei Abbruch/Recherche-Fehler. Default bei Lücke/Streit: 'relevant'."""
|
||||
topic, provider, is_cancelled = ctx.topic, ctx.provider, ctx.is_cancelled
|
||||
arbeit = files["arbeit"]
|
||||
items = [(titel, sub["titel"]) for titel, subs in sidecar.items() for sub in subs] # globale id = index+1
|
||||
if not items:
|
||||
return {}
|
||||
chunks = _chunk_nums(list(range(len(items))), _n_chunks(len(items), STUFE_CHUNK))
|
||||
n = len(chunks)
|
||||
|
||||
def rater_paths(c):
|
||||
return [arbeit / f"relevanz-c{c}-{i}.json" for i in (1, 2, 3)]
|
||||
|
||||
def lset(item_idxs):
|
||||
return set(range(1, len(item_idxs) + 1))
|
||||
|
||||
# Phase „Relevanz finden": pro Paket 3 Rater (min. 2), lokale IDs.
|
||||
async def _rate(c, item_idxs):
|
||||
local_set = lset(item_idxs)
|
||||
paths = rater_paths(c)
|
||||
vorhanden = sum(1 for p in paths if _relevanz_schema(_json_datei(p), local_set))
|
||||
if vorhanden >= 2:
|
||||
return True
|
||||
enum = "\n".join(f"{k}. [{items[j][0]}] {items[j][1]}" for k, j in enumerate(item_idxs, 1))
|
||||
offen = [(i, p) for i, p in enumerate(paths, 1) if not _relevanz_schema(_json_datei(p), local_set)]
|
||||
slots = [{
|
||||
"key": f"bausteine-{topic}-relevanz-c{c}-{i}",
|
||||
"prompt": _prompt("Relevanz-Recherche", topic=topic, subbausteine=enum, out_path=p, extra=_extra(instructions)),
|
||||
"role": "fast", "capabilities": "files",
|
||||
"payload": (lambda result, p=p, ids=local_set: _relevanz_schema(_json_datei(p), ids)),
|
||||
} for i, p in offen]
|
||||
neu = await _race(topic, f"Relevanz Paket {c}", slots, 2 - vorhanden, _timeout("relevanz", len(item_idxs)), provider, cancelled=is_cancelled, grace=KONSENS_GRACE)
|
||||
return not is_cancelled() and neu is not None
|
||||
|
||||
set_p(f"Relevanz finden ({n} Pakete)…", step=_step_idx(topic, "Relevanz finden"))
|
||||
oks = await asyncio.gather(*[_rate(c, idxs) for c, idxs in enumerate(chunks, 1)], return_exceptions=True)
|
||||
if is_cancelled():
|
||||
return None
|
||||
if not all(ok is True for ok in oks):
|
||||
_bausteine_errors[topic] = "Relevanz fehlgeschlagen (Recherche)"
|
||||
return None
|
||||
|
||||
# Phase „Relevanz wählen": Code-Vote je Paket → (ergebnis, strittig).
|
||||
set_p(f"Relevanz wählen ({n} Pakete)…", step=_step_idx(topic, "Relevanz wählen"))
|
||||
vote_by_c = {}
|
||||
for c, item_idxs in enumerate(chunks, 1):
|
||||
local_set = lset(item_idxs)
|
||||
rater = [d for p in rater_paths(c) if (d := _relevanz_schema(_json_datei(p), local_set))]
|
||||
ergebnis: dict[int, str] = {}
|
||||
strittig: dict[int, list[str]] = {}
|
||||
for k in range(1, len(item_idxs) + 1):
|
||||
stimmen = [d[k] for d in rater if k in d]
|
||||
zaehler: dict[str, int] = {}
|
||||
for s in stimmen:
|
||||
zaehler[s] = zaehler.get(s, 0) + 1
|
||||
best = max(zaehler.values(), default=0)
|
||||
gewinner = [s for s, v in zaehler.items() if v == best]
|
||||
if len(gewinner) == 1 and best >= 2:
|
||||
ergebnis[k] = gewinner[0]
|
||||
else:
|
||||
strittig[k] = stimmen
|
||||
vote_by_c[c] = (ergebnis, strittig)
|
||||
|
||||
# Phase „Relevanz klären": Judge je Paket mit Strittigem, alle parallel.
|
||||
async def _klaere(c, item_idxs):
|
||||
ergebnis, strittig = vote_by_c[c]
|
||||
if strittig:
|
||||
judge_path = arbeit / f"relevanz-final-c{c}.json"
|
||||
entsch = _relevanz_schema(_json_datei(judge_path), set(strittig))
|
||||
if entsch is None:
|
||||
strittig_block = "\n".join(
|
||||
f"{k}. [{items[item_idxs[k - 1]][0]}] {items[item_idxs[k - 1]][1]} — Stimmen: {', '.join(stimmen) or 'keine'}"
|
||||
for k, stimmen in strittig.items()
|
||||
)
|
||||
status, entsch = await run_single_slot(
|
||||
ctx, f"Relevanz-Klärung {c}",
|
||||
key=f"bausteine-{topic}-relevanz-final-c{c}",
|
||||
prompt=_prompt("Relevanz-Mapping", topic=topic, strittig=strittig_block, out_path=judge_path, extra=_extra(instructions)),
|
||||
role="judge", capabilities="files",
|
||||
payload=lambda result, p=judge_path, ids=set(strittig): _relevanz_schema(_json_datei(p), ids),
|
||||
timeout=_timeout("relevanz_check", len(strittig)),
|
||||
)
|
||||
if status == FAILED:
|
||||
_log(topic, f"Relevanz-Klärung Paket {c} fehlgeschlagen — Default 'relevant'")
|
||||
entsch = entsch if isinstance(entsch, dict) else {}
|
||||
# Strittige ohne Entscheid → 'relevant' (nie versehentlich ausschließen).
|
||||
ergebnis = {**{k: "relevant" for k in strittig}, **ergebnis, **entsch}
|
||||
return {item_idxs[k - 1] + 1: rel for k, rel in ergebnis.items()}
|
||||
|
||||
set_p(f"Relevanz klären ({n} Pakete)…", step=_step_idx(topic, "Relevanz klären"))
|
||||
parts = await asyncio.gather(*[_klaere(c, idxs) for c, idxs in enumerate(chunks, 1)], return_exceptions=True)
|
||||
if is_cancelled():
|
||||
return None
|
||||
relevanz_by_id: dict[int, str] = {}
|
||||
for c, part in enumerate(parts, 1):
|
||||
if not isinstance(part, dict):
|
||||
# Klärung ist nicht fatal: Vote-Ergebnis + Default 'relevant' für Strittige.
|
||||
if isinstance(part, BaseException):
|
||||
_log(topic, f"Relevanz-Klärung Paket {c}: {type(part).__name__}: {part}")
|
||||
ergebnis, strittig = vote_by_c[c]
|
||||
item_idxs = chunks[c - 1]
|
||||
merged = {**{k: "relevant" for k in strittig}, **ergebnis}
|
||||
part = {item_idxs[k - 1] + 1: s for k, s in merged.items()}
|
||||
relevanz_by_id.update(part)
|
||||
return relevanz_by_id
|
||||
|
||||
|
||||
async def generate_bausteine(topic: str, instructions: str = "", provider: str = DEFAULT_PROVIDER) -> None:
|
||||
if topic in _bausteine_progress:
|
||||
return
|
||||
@@ -834,6 +990,24 @@ async def generate_bausteine(topic: str, instructions: str = "", provider: str =
|
||||
if sidecar is None:
|
||||
return
|
||||
atomic_write_json(files["sidecar"], sidecar, indent=1)
|
||||
|
||||
# Block D: Relevanz je Subbaustein (relevant/rand) → in die Sidecar mergen.
|
||||
# Eigene Phase nach den Stufen; treibt das ProGuide-Format (alle Bausteine
|
||||
# mit ≥1 relevantem Subbaustein) und filtert Rand-Subs aus den Guides.
|
||||
sidecar = _json_datei(files["sidecar"])
|
||||
if _sidecar_schema(sidecar) is not None and not _relevanz_komplett(sidecar):
|
||||
relevanz_by_id = await _relevanz_block(ctx, set_p, files, sidecar, instructions)
|
||||
if is_cancelled():
|
||||
abgebrochen()
|
||||
return
|
||||
if relevanz_by_id is None:
|
||||
return # Fehler ist gesetzt
|
||||
gid = 0
|
||||
for subs in sidecar.values():
|
||||
for sub in subs:
|
||||
gid += 1
|
||||
sub["relevanz"] = relevanz_by_id.get(gid, "relevant")
|
||||
atomic_write_json(files["sidecar"], sidecar, indent=1)
|
||||
except Exception as e:
|
||||
log.exception("[%s] Bausteine-Generierung fehlgeschlagen", topic)
|
||||
_bausteine_errors[topic] = str(e)[:2000]
|
||||
|
||||
@@ -6,6 +6,7 @@ STORAGE_DIR = PROJECT_ROOT / "storage"
|
||||
FRONTEND_DIST = PROJECT_ROOT / "frontend" / "dist"
|
||||
DB_PATH = STORAGE_DIR / "creator.db"
|
||||
PROJECTS_DIR = PROJECT_ROOT / "projects"
|
||||
UNI_DIR = PROJECT_ROOT / "uni"
|
||||
|
||||
MAX_CONCURRENT_GENERATIONS = 10
|
||||
|
||||
@@ -42,6 +43,8 @@ TIMEOUTS = {
|
||||
"subbaustein_check": (300, 15), # Judge entscheidet strittige Subbausteine im Chunk
|
||||
"stufe": (300, 10), # Subbausteine einstufen je Chunk
|
||||
"stufe_check": (300, 10), # Judge entscheidet strittige Stufen im Chunk
|
||||
"relevanz": (300, 10), # Subbausteine relevant/rand je Chunk
|
||||
"relevanz_check": (300, 10), # Judge entscheidet strittige Relevanz im Chunk
|
||||
"writer": (600, 120), # pro Section im Chunk
|
||||
"lese_check": (300, 10), # pro Section im Paket
|
||||
"onepager_recherche": (900, 0),
|
||||
@@ -55,6 +58,7 @@ TIMEOUTS = {
|
||||
FORMAT_ANTEIL = {
|
||||
"MiniGuide": (0.05, 0.10, 8, "einen kompakten Anfänger-Guide — der schnelle Einstieg ins Thema"),
|
||||
"Guide": (0.25, 0.35, 20, "einen ausführlichen Anfänger-Guide — ein solides Fundament im Thema"),
|
||||
"ProGuide": (0.0, 1.0, 0, "einen fokussierten Guide — alles Relevante ohne Randthemen"),
|
||||
"FullGuide": (0.90, 1.00, 0, "einen Komplett-Guide — das ganze Thema"),
|
||||
}
|
||||
|
||||
|
||||
@@ -71,7 +71,7 @@ def _topic_context(topic: str, limit: int = 12000) -> str:
|
||||
bp = bausteine_path(topic)
|
||||
if bp.exists():
|
||||
parts.append(bp.read_text(encoding="utf-8"))
|
||||
for fmt in ("FullGuide", "Guide", "MiniGuide", "OnePager"):
|
||||
for fmt in ("FullGuide", "ProGuide", "Guide", "MiniGuide", "OnePager"):
|
||||
content = _json_datei(guide_content_path(topic, fmt))
|
||||
if content:
|
||||
for ch in content.get("chapters", []):
|
||||
|
||||
@@ -407,13 +407,26 @@ async def _generate_sections(
|
||||
"Wähle, was diesem Zweck dient — lass weg, was dafür nicht nötig ist."
|
||||
)
|
||||
|
||||
# Schritt 0: Auswahl-Konsens (5 Agenten → Voting → Mapping → Klärungs-Loop)
|
||||
auswahl = await _konsens_auswahl(
|
||||
ctx, files, entries, k_min, k_max, auswahl_auftrag, format_name,
|
||||
bausteine_liste, instructions,
|
||||
)
|
||||
if auswahl is None:
|
||||
return None
|
||||
# Subbausteine je Baustein (Sidecar) — früh geladen: ProGuide wählt darüber aus,
|
||||
# Nicht-FullGuide filtert später Rand-Subbausteine. Fehlt sie → {} (Fallback).
|
||||
subs_raw = _load_subbausteine(topic)
|
||||
|
||||
# Schritt 0: Auswahl. ProGuide ist deterministisch — alle Bausteine mit ≥1
|
||||
# relevantem Subbaustein. Fallback (keine Relevanz-Daten): alle Bausteine.
|
||||
if format_name == "ProGuide":
|
||||
auswahl = [num for num in entries
|
||||
if any(s.get("relevanz") == "relevant" for s in subs_raw.get(_titel(entries[num]), []))]
|
||||
if not auswahl:
|
||||
_log(topic, "ProGuide: keine Relevanz-Daten — alle Bausteine genommen")
|
||||
auswahl = list(entries)
|
||||
else:
|
||||
# Auswahl-Konsens (5 Agenten → Voting → Mapping → Klärungs-Loop)
|
||||
auswahl = await _konsens_auswahl(
|
||||
ctx, files, entries, k_min, k_max, auswahl_auftrag, format_name,
|
||||
bausteine_liste, instructions,
|
||||
)
|
||||
if auswahl is None:
|
||||
return None
|
||||
|
||||
sel_entries = {num: entries[num] for num in auswahl}
|
||||
soll = len(sel_entries)
|
||||
@@ -487,9 +500,16 @@ async def _generate_sections(
|
||||
# dieselbe Aufteilung: 1 Agent je ~30 Bausteine (gedeckelt).
|
||||
total_sections = sum(len(c["nums"]) for c in plan)
|
||||
chunks = _split_chunks(plan, min(WRITER_MAX, max(1, math.ceil(total_sections / WRITER_SECTIONS))))
|
||||
# Subbausteine je Baustein (Sidecar). Fehlt sie → leer → Inhalte/Writer laufen
|
||||
# auf Baustein-Granularität weiter (Fallback für Alt-Themen ohne Sidecar).
|
||||
subs_by_titel = _load_subbausteine(topic)
|
||||
# Subbausteine je Baustein. Rand-Subbausteine fliegen aus allen Guides AUSSER
|
||||
# FullGuide (Fokus). Subs ohne Relevanz-Feld (Alt-Themen) bleiben erhalten.
|
||||
# Fehlt die Sidecar → leer → Inhalte/Writer laufen auf Baustein-Granularität.
|
||||
subs_by_titel = subs_raw
|
||||
if format_name != "FullGuide":
|
||||
subs_by_titel = {
|
||||
t: [s for s in subs if s.get("relevanz") != "rand"]
|
||||
for t, subs in subs_raw.items()
|
||||
}
|
||||
subs_by_titel = {t: subs for t, subs in subs_by_titel.items() if subs}
|
||||
zuteilungen = [_zuteilung_subs(chunk, entries, subs_by_titel) for chunk in chunks]
|
||||
chunk_sizes = [sum(len(c["nums"]) for c in chunk) for chunk in chunks]
|
||||
writer_count = len(chunks)
|
||||
|
||||
@@ -5,6 +5,7 @@ FormatType = Literal[
|
||||
"OnePager",
|
||||
"MiniGuide",
|
||||
"Guide",
|
||||
"ProGuide",
|
||||
"FullGuide",
|
||||
]
|
||||
|
||||
@@ -50,6 +51,37 @@ class ProjectResponse(BaseModel):
|
||||
name: str
|
||||
|
||||
|
||||
class FolderResponse(BaseModel):
|
||||
name: str
|
||||
ort: str # relativer Pfad ab Repo-Root (z.B. "projects/foo")
|
||||
|
||||
|
||||
class BausteineQuelleUpdate(BaseModel):
|
||||
topic: str = Field(min_length=1, max_length=100)
|
||||
type: SourceType = "thema"
|
||||
ort: str = Field(default="", max_length=2000)
|
||||
spec: str = Field(default="", max_length=2000)
|
||||
|
||||
|
||||
class BausteineQuelleResponse(BaseModel):
|
||||
type: SourceType
|
||||
ort: str
|
||||
spec: str
|
||||
|
||||
|
||||
class SubbausteinInfo(BaseModel):
|
||||
titel: str
|
||||
stufe: Literal["einfach", "mittel", "schwer"]
|
||||
relevanz: Literal["relevant", "rand"] | None = None
|
||||
|
||||
|
||||
class BausteinUebersicht(BaseModel):
|
||||
num: int
|
||||
titel: str
|
||||
beschreibung: str = ""
|
||||
subbausteine: list[SubbausteinInfo] = []
|
||||
|
||||
|
||||
class ProviderInfo(BaseModel):
|
||||
id: str
|
||||
available: bool
|
||||
|
||||
@@ -158,6 +158,32 @@ def _stufen_schema(data, ids: set[int] | None = None):
|
||||
return out
|
||||
|
||||
|
||||
_RELEVANZ = ("relevant", "rand")
|
||||
|
||||
|
||||
def _relevanz_schema(data, ids: set[int] | None = None):
|
||||
"""{"relevanz": {"1": "relevant", …}} → {id: relevanz} · sonst None.
|
||||
|
||||
Relevanz ∈ {relevant, rand} (binär). Wie `_stufen_schema`: sind `ids` gegeben,
|
||||
müssen mindestens diese abgedeckt sein (Extras erlaubt).
|
||||
"""
|
||||
if not isinstance(data, dict) or not isinstance(data.get("relevanz"), dict) or not data["relevanz"]:
|
||||
return None
|
||||
out: dict[int, str] = {}
|
||||
for k, v in data["relevanz"].items():
|
||||
try:
|
||||
num = int(k)
|
||||
except (ValueError, TypeError):
|
||||
return None
|
||||
wert = str(v).strip().casefold()
|
||||
if wert not in _RELEVANZ:
|
||||
return None
|
||||
out[num] = wert
|
||||
if ids is not None and not ids <= set(out):
|
||||
return None
|
||||
return out
|
||||
|
||||
|
||||
_MAX_RESTARTS = 2
|
||||
|
||||
|
||||
|
||||
@@ -18,10 +18,11 @@ from paths import bausteine_path, guide_content_path
|
||||
from textkit import _norm_titel
|
||||
|
||||
MAX_OFFENE_GUIDES = 3
|
||||
VORSTUFE = {"Guide": "MiniGuide", "FullGuide": "Guide"}
|
||||
VORSTUFE = {"Guide": "MiniGuide", "ProGuide": "Guide", "FullGuide": "Guide"}
|
||||
# Welches Niveau die Vorstufe erreichen muss, um das Format freizuschalten.
|
||||
FREISCHALT_LEVEL = {"Guide": "absolviert", "FullGuide": "verstanden"}
|
||||
FORMATE = ("MiniGuide", "Guide", "FullGuide")
|
||||
# ProGuide hängt parallel zu FullGuide an „Guide absolviert" — kein Pflicht-Gate vor FullGuide.
|
||||
FREISCHALT_LEVEL = {"Guide": "absolviert", "ProGuide": "absolviert", "FullGuide": "verstanden"}
|
||||
FORMATE = ("MiniGuide", "Guide", "ProGuide", "FullGuide")
|
||||
|
||||
|
||||
async def lade_lernstand() -> tuple[list[dict], dict[str, set[str]], dict[str, dict[str, set[str]]]]:
|
||||
|
||||
@@ -7,7 +7,7 @@ from fastapi import APIRouter, HTTPException
|
||||
from fastapi.responses import FileResponse
|
||||
|
||||
from agents import provider_available
|
||||
from config import PROJECTS_DIR, PROVIDERS
|
||||
from config import PROJECTS_DIR, UNI_DIR, PROVIDERS
|
||||
from database import (
|
||||
create_guide, delete_guide, get_guide, list_guides,
|
||||
create_topic, list_topics as db_list_topics, delete_topic,
|
||||
@@ -16,7 +16,7 @@ from database import (
|
||||
get_vertiefung, set_vertiefung, list_vertiefungen,
|
||||
list_baustein_progress, set_baustein_score, set_baustein_streak, set_baustein_absolviert, set_baustein_verstanden, set_baustein_gemeistert, delete_baustein_daten,
|
||||
)
|
||||
from bausteine import generate_bausteine, cancel_bausteine, bausteine_status, active_bausteine, reset_bausteine
|
||||
from bausteine import generate_bausteine, cancel_bausteine, bausteine_status, active_bausteine, reset_bausteine, lade_quelle, lade_uebersicht
|
||||
from elements import generate_element, chat_with_guide, chat_with_element, check_element, style_element, refine_suggestion
|
||||
from lernen import NOETIG, MASTERY, MEISTERN, baustein_chat, baustein_diskussion, baustein_element_anlegen, pruefung_bewertung, pruefung_bewertung_schnell, pruefung_frage, score_berechnen, streak_berechnen, vertiefung_generieren
|
||||
from guide import generate_guide, guide_slot_dateien
|
||||
@@ -31,6 +31,7 @@ from models import (
|
||||
ElementUpdateRequest, ElementCheckRequest, ElementCheckResponse, ElementStyleResponse,
|
||||
ElementRefineRequest, ElementRefineResponse,
|
||||
ProgressUpdate, ProgressResponse, ProjectResponse, ProviderInfo,
|
||||
FolderResponse, BausteineQuelleUpdate, BausteineQuelleResponse, BausteinUebersicht,
|
||||
VertiefungRequest, VertiefungResponse,
|
||||
BausteinChatRequest, BausteinChatResponse,
|
||||
BausteinPruefungRequest, BausteinPruefungResponse, BausteinLernstandResponse,
|
||||
@@ -113,6 +114,18 @@ async def remove_project(name: str):
|
||||
return {"ok": True}
|
||||
|
||||
|
||||
@router.get("/folders", response_model=list[FolderResponse])
|
||||
async def list_folders(kind: str):
|
||||
"""Ordner für die Quellen-Auswahl: kind=projekt → projects/, kind=uni → uni/."""
|
||||
base = {"projekt": (PROJECTS_DIR, "projects"), "uni": (UNI_DIR, "uni")}.get(kind)
|
||||
if base is None:
|
||||
raise HTTPException(400, "kind muss 'projekt' oder 'uni' sein")
|
||||
root, prefix = base
|
||||
if not root.is_dir():
|
||||
return []
|
||||
return [{"name": e.name, "ort": f"{prefix}/{e.name}"} for e in sorted(root.iterdir()) if e.is_dir()]
|
||||
|
||||
|
||||
# --- Bausteine ---
|
||||
|
||||
@router.get("/bausteine/status", response_model=BausteineStatusResponse)
|
||||
@@ -161,6 +174,39 @@ async def remove_bausteine(topic: str):
|
||||
return {"ok": True}
|
||||
|
||||
|
||||
def _validate_quelle(typ: str, ort: str) -> None:
|
||||
"""Quellen-Eingabe prüfen (gleiche Regeln wie beim Erstellen)."""
|
||||
if typ in ("projekt", "uni"):
|
||||
ordner = safe_ordner(ort)
|
||||
if ordner is None or not ordner.is_dir():
|
||||
raise HTTPException(400, "Ordner ungültig oder nicht gefunden (Pfad relativ zum Projekt-Root, kein ../).")
|
||||
elif typ == "link":
|
||||
if not ort.lower().startswith(("http://", "https://")):
|
||||
raise HTTPException(400, "Link muss mit http:// oder https:// beginnen.")
|
||||
|
||||
|
||||
@router.get("/bausteine/quelle", response_model=BausteineQuelleResponse)
|
||||
async def get_bausteine_quelle(topic: str):
|
||||
return lade_quelle(topic)
|
||||
|
||||
|
||||
@router.put("/bausteine/quelle", response_model=BausteineQuelleResponse)
|
||||
async def update_bausteine_quelle(req: BausteineQuelleUpdate):
|
||||
"""Nur speichern — KEINE Neugenerierung. Quellen-/Spec-Wahl überschreiben."""
|
||||
topic, typ, ort = req.topic.strip(), req.type, req.ort.strip()
|
||||
_validate_quelle(typ, ort)
|
||||
qp = quelle_path(topic)
|
||||
qp.parent.mkdir(parents=True, exist_ok=True)
|
||||
daten = {"type": typ, "ort": ort, "spec": req.spec.strip()}
|
||||
atomic_write_json(qp, daten)
|
||||
return daten
|
||||
|
||||
|
||||
@router.get("/bausteine/uebersicht", response_model=list[BausteinUebersicht])
|
||||
async def get_bausteine_uebersicht(topic: str):
|
||||
return lade_uebersicht(topic)
|
||||
|
||||
|
||||
# --- Baustein-Lernen: Vertiefung, Chat, Prüfung ---
|
||||
|
||||
async def _bester_text(topic: str, baustein: str) -> str | None:
|
||||
|
||||
Reference in New Issue
Block a user