mirror of
https://github.com/basicmachines-co/basic-memory
synced 2026-06-21 13:47:35 +00:00
split up entity parsing
This commit is contained in:
phernandez
@@ -3,12 +3,8 @@
|
||||
import logging
|
||||
from pathlib import Path
|
||||
|
||||
from markdown_it import MarkdownIt
|
||||
|
||||
from basic_memory.markdown.exceptions import ParseError
|
||||
from basic_memory.markdown.schemas import (
|
||||
Observation,
|
||||
Relation,
|
||||
Entity,
|
||||
EntityFrontmatter,
|
||||
EntityContent,
|
||||
@@ -31,9 +27,6 @@ def debug_sections(text):
|
||||
class EntityParser:
|
||||
"""Parser for entity markdown files."""
|
||||
|
||||
def __init__(self):
|
||||
self.md = MarkdownIt()
|
||||
|
||||
def parse_file(self, path: Path, encoding: str = "utf-8") -> Entity:
|
||||
"""Parse an entity markdown file."""
|
||||
if not path.exists():
|
||||
@@ -50,80 +43,10 @@ class EntityParser:
|
||||
|
||||
# Parse each section using schema methods
|
||||
frontmatter = EntityFrontmatter.from_text(sections[1])
|
||||
content = EntityContent.from_markdown(sections[2])
|
||||
metadata = EntityMetadata.from_text(sections[4] if len(sections) >= 5 else "")
|
||||
|
||||
# Parse markdown content (middle section)
|
||||
content_tokens = self.md.parse(sections[2].strip())
|
||||
|
||||
# State for content parsing
|
||||
title = ""
|
||||
description = ""
|
||||
observations = []
|
||||
relations = []
|
||||
current_section = None
|
||||
|
||||
# Track list items
|
||||
in_list_item = False
|
||||
list_item_tokens = []
|
||||
|
||||
for token in content_tokens:
|
||||
if token.type == "heading_open":
|
||||
if token.tag == "h1":
|
||||
current_section = "title"
|
||||
elif token.tag == "h2":
|
||||
current_section = "section_name"
|
||||
|
||||
elif token.type == "inline":
|
||||
content = token.content.strip()
|
||||
|
||||
if current_section == "title":
|
||||
title = content
|
||||
current_section = "description"
|
||||
elif current_section == "section_name":
|
||||
current_section = content.lower()
|
||||
elif current_section == "description":
|
||||
if description:
|
||||
description += " "
|
||||
description += content
|
||||
elif in_list_item:
|
||||
list_item_tokens.append(token)
|
||||
|
||||
elif token.type == "list_item_open":
|
||||
in_list_item = True
|
||||
list_item_tokens = []
|
||||
|
||||
elif token.type == "list_item_close":
|
||||
item_content = " ".join(t.content for t in list_item_tokens)
|
||||
try:
|
||||
if current_section == "observations":
|
||||
if obs := Observation.from_line(item_content):
|
||||
observations.append(obs)
|
||||
elif current_section == "relations":
|
||||
if rel := Relation.from_line(item_content):
|
||||
relations.append(rel)
|
||||
except ParseError:
|
||||
# Skip malformed items
|
||||
pass
|
||||
in_list_item = False
|
||||
|
||||
# Create content object
|
||||
content = EntityContent(
|
||||
title=title,
|
||||
description=description,
|
||||
observations=observations,
|
||||
relations=relations,
|
||||
)
|
||||
|
||||
# Parse metadata from final section
|
||||
metadata_obj = EntityMetadata(metadata={})
|
||||
if len(sections) >= 5:
|
||||
metadata_text = sections[4].strip()
|
||||
logger.debug(f"Metadata text: {metadata_text}")
|
||||
for line in metadata_text.split("\n"):
|
||||
if ":" in line:
|
||||
key, value = line.split(":", 1)
|
||||
metadata_obj.metadata[key.strip()] = value.strip()
|
||||
|
||||
return Entity(frontmatter=frontmatter, content=content, metadata=metadata_obj)
|
||||
return Entity(frontmatter=frontmatter, content=content, metadata=metadata)
|
||||
|
||||
except UnicodeError as e:
|
||||
if encoding == "utf-8":
|
||||
|
||||
@@ -4,6 +4,7 @@ import logging
|
||||
from datetime import datetime
|
||||
from typing import Any, Dict, List, Optional
|
||||
|
||||
from markdown_it import MarkdownIt
|
||||
from pydantic import BaseModel
|
||||
|
||||
from basic_memory.markdown.exceptions import ParseError
|
||||
@@ -47,6 +48,20 @@ class EntityMetadata(BaseModel):
|
||||
|
||||
metadata: Dict[str, Any] = {}
|
||||
|
||||
@classmethod
|
||||
def from_text(cls, text: str) -> "EntityMetadata":
|
||||
"""Parse metadata from text."""
|
||||
try:
|
||||
metadata = {}
|
||||
if text: # Only parse if there's content
|
||||
for line in text.strip().split("\n"):
|
||||
if ":" in line:
|
||||
key, value = line.split(":", 1)
|
||||
metadata[key.strip()] = value.strip()
|
||||
return cls(metadata=metadata)
|
||||
except Exception as e:
|
||||
raise ParseError(f"Failed to parse metadata: {e}") from e
|
||||
|
||||
|
||||
class EntityContent(BaseModel):
|
||||
"""Content sections of an entity markdown file."""
|
||||
@@ -57,6 +72,74 @@ class EntityContent(BaseModel):
|
||||
relations: List[Relation] = []
|
||||
context: Optional[str] = None
|
||||
|
||||
@classmethod
|
||||
def from_markdown(cls, text: str) -> "EntityContent":
|
||||
"""Parse content from markdown text."""
|
||||
try:
|
||||
md = MarkdownIt()
|
||||
tokens = md.parse(text.strip())
|
||||
|
||||
# State for parsing
|
||||
title = ""
|
||||
description = ""
|
||||
observations: List[Observation] = []
|
||||
relations: List[Relation] = []
|
||||
current_section = None
|
||||
|
||||
# Track list items
|
||||
in_list_item = False
|
||||
list_item_tokens = []
|
||||
|
||||
for token in tokens:
|
||||
if token.type == "heading_open":
|
||||
if token.tag == "h1":
|
||||
current_section = "title"
|
||||
elif token.tag == "h2":
|
||||
current_section = "section_name"
|
||||
|
||||
elif token.type == "inline":
|
||||
content = token.content.strip()
|
||||
|
||||
if current_section == "title":
|
||||
title = content
|
||||
current_section = "description"
|
||||
elif current_section == "section_name":
|
||||
current_section = content.lower()
|
||||
elif current_section == "description":
|
||||
if description:
|
||||
description += " "
|
||||
description += content
|
||||
elif in_list_item:
|
||||
list_item_tokens.append(token)
|
||||
|
||||
elif token.type == "list_item_open":
|
||||
in_list_item = True
|
||||
list_item_tokens = []
|
||||
|
||||
elif token.type == "list_item_close":
|
||||
item_content = " ".join(t.content for t in list_item_tokens)
|
||||
try:
|
||||
if current_section == "observations":
|
||||
if obs := Observation.from_line(item_content):
|
||||
observations.append(obs)
|
||||
elif current_section == "relations":
|
||||
if rel := Relation.from_line(item_content):
|
||||
relations.append(rel)
|
||||
except ParseError:
|
||||
# Skip malformed items
|
||||
pass
|
||||
in_list_item = False
|
||||
|
||||
return cls(
|
||||
title=title,
|
||||
description=description,
|
||||
observations=observations,
|
||||
relations=relations,
|
||||
)
|
||||
|
||||
except Exception as e:
|
||||
raise ParseError(f"Failed to parse markdown content: {e}") from e
|
||||
|
||||
|
||||
class Entity(BaseModel):
|
||||
"""Complete entity combining frontmatter, content, and metadata."""
|
||||
|
||||
Reference in New Issue
Block a user