fabricatio 0.2.13.dev1__cp312-cp312-win_amd64.whl → 0.2.13.dev3__cp312-cp312-win_amd64.whl

fabricatio/actions/article_rag.py

@@ -20,18 +20,26 @@ from fabricatio.models.extra.article_main import Article, ArticleChapter, Articl
 from fabricatio.models.extra.article_outline import ArticleOutline
 from fabricatio.models.extra.rule import RuleSet
 from fabricatio.models.kwargs_types import ChooseKwargs, LLMKwargs
-from fabricatio.rust import convert_to_block_formula, convert_to_inline_formula
-from fabricatio.utils import ask_retain, ok
+from fabricatio.rust import (
+    convert_all_block_tex,
+    convert_all_inline_tex,
+    convert_to_block_formula,
+    convert_to_inline_formula,
+    fix_misplaced_labels,
+)
+from fabricatio.utils import ok
 
 TYPST_CITE_USAGE = (
-    "citation number is REQUIRED to cite any reference!,for example in Auther Pattern: 'Doe et al.[[1]], Jack et al.[[2]]' or in Sentence Suffix Sattern: 'Global requirement is incresing[[1]].'\n"
-    "Everything is build upon the typst language, which is similar to latex, \n"
+    "citation number is REQUIRED to cite any reference!'\n"
     "Legal citing syntax examples(seperated by |): [[1]]|[[1,2]]|[[1-3]]|[[12,13-15]]|[[1-3,5-7]]\n"
     "Illegal citing syntax examples(seperated by |): [[1],[2],[3]]|[[1],[1-2]]\n"
-    "Those reference mark shall not be omitted during the extraction\n"
+    "You SHALL not cite a single reference more than once!"
     "It's recommended to cite multiple references that supports your conclusion at a time.\n"
+)
+
+TYPST_MATH_USAGE = (
     "Wrap inline expression with '\\(' and '\\)',like '\\(>5m\\)' '\\(89%\\)', and wrap block equation with '\\[' and '\\]'.\n"
-    "In addition to that, you can add a label outside the block equation which can be used as a cross reference identifier, the label is a string wrapped in `<` and `>` like `<energy-release-rate-equation>`.Note that the label string should be a summarizing title for the equation being labeled.\n"
+    "In addition to that, you can add a label outside the block equation which can be used as a cross reference identifier, the label is a string wrapped in `<` and `>` like `<energy-release-rate-equation>`.Note that the label string should be a summarizing title for the equation being labeled and should never be written within the formula block.\n"
     "you can refer to that label by using the syntax with prefix of `@eqt:`, which indicate that this notation is citing a label from the equations. For example ' @eqt:energy-release-rate-equation ' DO remember that the notation shall have both suffixed and prefixed space char which enable the compiler to distinguish the notation from the plaintext."
     "Below is two usage example:\n"
     "```typst\n"
@@ -44,9 +52,10 @@ TYPST_CITE_USAGE = (
 )
 
 
-class WriteArticleContentRAG(Action, RAG, Extract):
+class WriteArticleContentRAG(Action, Extract, AdvancedRAG):
     """Write an article based on the provided outline."""
 
+    ctx_override: ClassVar[bool] = True
     search_increment_multiplier: float = 1.6
     """The increment multiplier of the search increment."""
     ref_limit: int = 35
@@ -55,15 +64,19 @@ class WriteArticleContentRAG(Action, RAG, Extract):
     """The threshold of relevance"""
     extractor_model: LLMKwargs
     """The model to use for extracting the content from the retrieved references."""
-    query_model: LLMKwargs
+    query_model: ChooseKwargs
     """The model to use for querying the database"""
     supervisor: bool = False
     """Whether to use supervisor mode"""
     result_per_query: PositiveInt = 4
     """The number of results to be returned per query."""
-    req: str = TYPST_CITE_USAGE
+    cite_req: str = TYPST_CITE_USAGE
     """The req of the write article content."""
 
+    math_req: str = TYPST_MATH_USAGE
+    """The req of the write article content."""
+    tei_endpoint: Optional[str] = None
+
     async def _execute(
         self,
         article_outline: ArticleOutline,
@@ -103,30 +116,25 @@ class WriteArticleContentRAG(Action, RAG, Extract):
         cm = CitationManager()
         await self.search_database(article, article_outline, chap, sec, subsec, cm)
 
-        raw = await self.write_raw(article, article_outline, chap, sec, subsec, cm)
-        r_print(raw)
+        raw_paras = await self.write_raw(article, article_outline, chap, sec, subsec, cm)
+        r_print(raw_paras)
 
         while not await confirm("Accept this version and continue?").ask_async():
             if inst := await text("Search for more refs for additional spec.").ask_async():
-                await self.search_database(
-                    article,
-                    article_outline,
-                    chap,
-                    sec,
-                    subsec,
-                    cm,
-                    supervisor=True,
-                    extra_instruction=inst,
-                )
+                await self.search_database(article, article_outline, chap, sec, subsec, cm, extra_instruction=inst)
 
             if instruction := await text("Enter the instructions to improve").ask_async():
-                raw = await self.write_raw(article, article_outline, chap, sec, subsec, cm, instruction)
-            if edt := await text("Edit", default=raw).ask_async():
-                raw = edt
+                raw_paras = await self.write_raw(article, article_outline, chap, sec, subsec, cm, instruction)
+            if edt := await text("Edit", default=raw_paras).ask_async():
+                raw_paras = edt
 
-            r_print(raw)
+            raw_paras = fix_misplaced_labels(raw_paras)
+            raw_paras = convert_all_inline_tex(raw_paras)
+            raw_paras = convert_all_block_tex(raw_paras)
 
-        return await self.extract_new_subsec(subsec, raw, cm)
+            r_print(raw_paras)
+
+        return await self.extract_new_subsec(subsec, raw_paras, cm)
 
     async def _inner(
         self,
@@ -144,6 +152,10 @@ class WriteArticleContentRAG(Action, RAG, Extract):
 
         raw_paras = "\n".join(p for p in raw_paras.splitlines() if p and not p.endswith("**") and not p.startswith("#"))
 
+        raw_paras = fix_misplaced_labels(raw_paras)
+        raw_paras = convert_all_inline_tex(raw_paras)
+        raw_paras = convert_all_block_tex(raw_paras)
+
         return await self.extract_new_subsec(subsec, raw_paras, cm)
 
     async def extract_new_subsec(
@@ -155,12 +167,13 @@ class WriteArticleContentRAG(Action, RAG, Extract):
                 ArticleSubsection,
                 raw_paras,
                 f"Above is the subsection titled `{subsec.title}`.\n"
-                f"I need you to extract the content to update my subsection obj provided below.\n{self.req}"
-                f"{subsec.display()}\n",
+                f"I need you to extract the content to construct a new `{ArticleSubsection.__class__.__name__}`,"
+                f"Do not attempt to change the original content, your job is ONLY content extraction",
                 **self.extractor_model,
             ),
             "Failed to propose new subsection.",
         )
+
         for p in new_subsec.paragraphs:
             p.content = cm.apply(p.content)
             p.description = cm.apply(p.description)
@@ -184,7 +197,7 @@ class WriteArticleContentRAG(Action, RAG, Extract):
             f"{article_outline.finalized_dump()}\n\nAbove is my article outline, I m writing graduate thesis titled `{article.title}`. "
             f"More specifically, i m witting the Chapter `{chap.title}` >> Section `{sec.title}` >> Subsection `{subsec.title}`.\n"
             f"Please help me write the paragraphs of the subsec mentioned above, which is `{subsec.title}`.\n"
-            f"{self.req}\n"
+            f"{self.cite_req}\n{self.math_req}\n"
             f"You SHALL use `{article.language}` as writing language.\n{extra_instruction}\n"
             f"Do not use numbered list to display the outcome, you should regard you are writing the main text of the thesis.\n"
             f"You should not copy others' works from the references directly on to my thesis, we can only harness the conclusion they have drawn.\n"
@@ -200,7 +213,6 @@ class WriteArticleContentRAG(Action, RAG, Extract):
         subsec: ArticleSubsection,
         cm: CitationManager,
         extra_instruction: str = "",
-        supervisor: bool = False,
     ) -> None:
         """Search database for related references."""
         search_req = (
@@ -208,51 +220,20 @@ class WriteArticleContentRAG(Action, RAG, Extract):
             f"More specifically, i m witting the Chapter `{chap.title}` >> Section `{sec.title}` >> Subsection `{subsec.title}`.\n"
             f"I need to search related references to build up the content of the subsec mentioned above, which is `{subsec.title}`.\n"
             f"provide 10~16 queries as possible, to get best result!\n"
-            f"You should provide both English version and chinese version of the refined queries!\n{extra_instruction}\n"
-        )
-
-        ref_q = ok(
-            await self.arefined_query(
-                search_req,
-                **self.query_model,
-            ),
-            "Failed to refine query.",
+            f"You should provide both English version and chinese version of the refined queries!\n{extra_instruction}"
         )
 
-        if supervisor:
-            ref_q = await ask_retain(ref_q)
-        ret = await self.aretrieve(
-            ref_q,
-            ArticleChunk,
-            max_accepted=self.ref_limit,
+        await self.clued_search(
+            search_req,
+            cm,
+            refinery_kwargs=self.query_model,
+            expand_multiplier=self.search_increment_multiplier,
+            base_accepted=self.ref_limit,
             result_per_query=self.result_per_query,
             similarity_threshold=self.threshold,
+            tei_endpoint=self.tei_endpoint,
         )
 
-        cm.add_chunks(ok(ret))
-        ref_q = await self.arefined_query(
-            f"{cm.as_prompt()}\n\nAbove is the retrieved references in the first RAG, now we need to perform the second RAG.\n\n{search_req}",
-            **self.query_model,
-        )
-
-        if ref_q is None:
-            logger.warning("Second refine query is None, skipping.")
-            return
-        if supervisor:
-            ref_q = await ask_retain(ref_q)
-
-        ret = await self.aretrieve(
-            ref_q,
-            ArticleChunk,
-            max_accepted=int(self.ref_limit * self.search_increment_multiplier),
-            result_per_query=int(self.result_per_query * self.search_increment_multiplier),
-            similarity_threshold=self.threshold,
-        )
-        if ret is None:
-            logger.warning("Second retrieve is None, skipping.")
-            return
-        cm.add_chunks(ret)
-
 
 class ArticleConsultRAG(Action, AdvancedRAG):
     """Write an article based on the provided outline."""
@@ -261,9 +242,9 @@ class ArticleConsultRAG(Action, AdvancedRAG):
     output_key: str = "consult_count"
     search_increment_multiplier: float = 1.6
     """The multiplier to increase the limit of references to retrieve per query."""
-    ref_limit: int = 20
+    ref_limit: int = 26
     """The final limit of references."""
-    ref_per_q: int = 3
+    ref_per_q: int = 13
     """The limit of references to retrieve per query."""
     similarity_threshold: float = 0.62
     """The similarity threshold of references to retrieve."""

fabricatio/capabilities/advanced_rag.py

@@ -7,6 +7,7 @@ from fabricatio.journal import logger
 from fabricatio.models.adv_kwargs_types import FetchKwargs
 from fabricatio.models.extra.aricle_rag import ArticleChunk, CitationManager
 from fabricatio.models.kwargs_types import ChooseKwargs
+from fabricatio.utils import fallback_kwargs
 
 
 class AdvancedRAG(RAG):
@@ -40,10 +41,13 @@ class AdvancedRAG(RAG):
                 f"\n\n{requirement}",
                 **refinery_kwargs,
             )
+
             if ref_q is None:
                 logger.error(f"At round [{i}/{max_round}] search, failed to refine the query, exit.")
                 return cm
-            refs = await self.aretrieve(ref_q, ArticleChunk, base_accepted, **kwargs)
+            refs = await self.aretrieve(
+                ref_q, ArticleChunk, base_accepted, **fallback_kwargs(kwargs, filter_expr=cm.as_milvus_filter_expr())
+            )
 
             if (max_capacity := max_capacity - len(refs)) < 0:
                 cm.add_chunks(refs[0:max_capacity])

fabricatio/capabilities/rag.py

@@ -150,6 +150,7 @@ class RAG(EmbeddingUsage):
         result_per_query: int = 10,
         tei_endpoint: Optional[str] = None,
         reranker_threshold: float = 0.7,
+        filter_expr: str = "",
     ) -> List[D]:
         """Asynchronously fetches documents from a Milvus database based on input vectors.
 
@@ -162,6 +163,7 @@ class RAG(EmbeddingUsage):
            result_per_query (int): The maximum number of results to return per query. Defaults to 10.
            tei_endpoint (str): the endpoint of the TEI api.
            reranker_threshold (float): The threshold used to filtered low relativity document.
+            filter_expr (str): filter_expression parsed into pymilvus search.
 
         Returns:
            List[D]: A list of document objects created from the fetched data.
@@ -172,6 +174,7 @@ class RAG(EmbeddingUsage):
             await self.vectorize(query),
             search_params={"radius": similarity_threshold},
             output_fields=list(document_model.model_fields),
+            filter=filter_expr,
             limit=result_per_query,
         )
         if tei_endpoint is not None:
@@ -184,8 +187,11 @@ class RAG(EmbeddingUsage):
 
             for q, g in zip(query, search_results, strict=True):
                 models = document_model.from_sequence([res["entity"] for res in g if res["id"] not in retrieved_id])
+                logger.debug(f"Retrived {len(g)} raw document, filtered out {len(models)}.")
                 retrieved_id.update(res["id"] for res in g)
-                rank_scores = await reranker.arerank(q, [m.prepare_vectorization() for m in models])
+                if not models:
+                    continue
+                rank_scores = await reranker.arerank(q, [m.prepare_vectorization() for m in models], truncate=True)
                 raw_result.extend(
                     (models[s["index"]], s["score"]) for s in rank_scores if s["score"] > reranker_threshold
                 )

fabricatio/models/adv_kwargs_types.py

@@ -60,3 +60,4 @@ if find_spec("pymilvus"):
         result_per_query: NotRequired[int]
         tei_endpoint: NotRequired[Optional[str]]
         reranker_threshold: NotRequired[float]
+        filter_expr: NotRequired[str]

fabricatio/models/extra/aricle_rag.py

@@ -1,6 +1,7 @@
 """A Module containing the article rag models."""
 
 import re
+from itertools import groupby
 from pathlib import Path
 from typing import ClassVar, Dict, List, Optional, Self, Unpack
 
@@ -10,12 +11,13 @@ from fabricatio.models.extra.rag import MilvusDataBase
 from fabricatio.models.generic import AsPrompt
 from fabricatio.models.kwargs_types import ChunkKwargs
 from fabricatio.rust import BibManager, blake3_hash, split_into_chunks
-from fabricatio.utils import ok
+from fabricatio.utils import ok, wrapp_in_block
+from more_itertools.more import first
 from more_itertools.recipes import flatten, unique
 from pydantic import Field
 
 
-class ArticleChunk(MilvusDataBase, AsPrompt):
+class ArticleChunk(MilvusDataBase):
     """The chunk of an article."""
 
     etc_word: ClassVar[str] = "等"
@@ -51,10 +53,9 @@ class ArticleChunk(MilvusDataBase, AsPrompt):
     bibtex_cite_key: str
     """The bibtex cite key of the article"""
 
-    def _as_prompt_inner(self) -> Dict[str, str]:
-        return {
-            f"[[{ok(self._cite_number, 'You need to update cite number first.')}]] reference `{self.article_title}` from {self.as_auther_seq()}": self.chunk
-        }
+    @property
+    def reference_header(self) -> str:
+        return f"[[{ok(self._cite_number, 'You need to update cite number first.')}]] reference `{self.article_title}` from {self.as_auther_seq()}"
 
     @property
     def cite_number(self) -> int:
@@ -204,13 +205,23 @@ class CitationManager(AsPrompt):
 
     def set_cite_number_all(self) -> Self:
         """Set citation numbers for all article chunks."""
-        for i, a in enumerate(self.article_chunks, 1):
-            a.update_cite_number(i)
+        number_mapping = {a.bibtex_cite_key: 0 for a in self.article_chunks}
+
+        for i, k in enumerate(number_mapping.keys()):
+            number_mapping[k] = i
+
+        for a in self.article_chunks:
+            a.update_cite_number(number_mapping[a.bibtex_cite_key])
         return self
 
     def _as_prompt_inner(self) -> Dict[str, str]:
         """Generate prompt inner representation."""
-        return {"References": "\n".join(r.as_prompt() for r in self.article_chunks)}
+        seg = []
+        for k, g in groupby(self.article_chunks, key=lambda a: a.bibtex_cite_key):
+            g = list(g)
+            logger.debug(f"Group [{k}]: {len(g)}")
+            seg.append(wrapp_in_block("\n\n".join(a.chunk for a in g), first(g).reference_header))
+        return {"References": "\n".join(seg)}
 
     def apply(self, string: str) -> str:
         """Apply citation replacements to the input string."""
@@ -261,5 +272,9 @@ class CitationManager(AsPrompt):
 
     def unpack_cite_seq(self, citation_seq: List[int]) -> str:
         """Unpack citation sequence into a string."""
-        chunk_seq = [a for a in self.article_chunks if a.cite_number in citation_seq]
-        return "".join(a.as_typst_cite() for a in chunk_seq)
+        chunk_seq = {a.bibtex_cite_key: a for a in self.article_chunks if a.cite_number in citation_seq}
+        return "".join(a.as_typst_cite() for a in chunk_seq.values())
+
+    def as_milvus_filter_expr(self, blacklist: bool = True) -> str:
+        if blacklist:
+            return " and ".join(f'bibtex_cite_key != "{a.bibtex_cite_key}"' for a in self.article_chunks)

fabricatio/models/extra/article_base.py

@@ -2,9 +2,12 @@
 
 from abc import ABC
 from enum import StrEnum
+from pathlib import Path
 from typing import ClassVar, Generator, List, Optional, Self, Tuple, Type
 
+from fabricatio.fs import dump_text, safe_text_read
 from fabricatio.fs.readers import extract_sections
+from fabricatio.journal import logger
 from fabricatio.models.generic import (
     AsPrompt,
     Described,
@@ -19,10 +22,12 @@ from fabricatio.models.generic import (
     Titled,
     WordCount,
 )
-from fabricatio.rust import split_out_metadata, to_metadata, word_count
-from fabricatio.utils import fallback_kwargs
+from fabricatio.rust import extract_body, inplace_update, split_out_metadata, to_metadata, word_count
+from fabricatio.utils import fallback_kwargs, ok
 from pydantic import Field
 
+ARTICLE_WRAPPER = "// =-=-=-=-=-=-=-=-=-="
+
 
 class ReferringType(StrEnum):
     """Enumeration of different types of references that can be made in an article."""
@@ -397,3 +402,23 @@ class ArticleBase[T: ChapterBase](FinalizedDumpAble, AsPrompt, FromTypstCode, To
     def avg_wordcount_recursive[S: "ArticleBase"](self: S) -> S:
         """Set all chap, sec, subsec have same word count sum up to be `self.expected_word_count`."""
         return self.avg_chap_wordcount().avg_sec_wordcount().avg_subsec_wordcount()
+
+    def update_article_file(self, file: str | Path) -> Self:
+        """Update the article file."""
+        file = Path(file)
+        string = safe_text_read(file)
+        if updated := inplace_update(string, ARTICLE_WRAPPER, self.to_typst_code()):
+            dump_text(file, updated)
+            logger.success(f"Successfully updated {file.as_posix()}.")
+        else:
+            logger.warning(f"Failed to update {file.as_posix()}. Please make sure there are paired `{ARTICLE_WRAPPER}`")
+        return self
+
+    @classmethod
+    def from_article_file(cls, file: str | Path, title: str) -> Self:
+        """Load article from file."""
+        file = Path(file)
+        string = safe_text_read(file)
+        return cls.from_typst_code(
+            title, ok(extract_body(string, ARTICLE_WRAPPER), "Failed to extract body from file.")
+        )

fabricatio/models/extra/article_main.py

@@ -53,6 +53,7 @@ class Paragraph(SketchedAble, WordCount, Described):
 
     @property
     def exact_wordcount(self) -> int:
+        """Calculates the exact word count of the content."""
         return word_count(self.content)
 
 

fabricatio/models/extra/article_outline.py

@@ -1,8 +1,7 @@
 """A module containing the ArticleOutline class, which represents the outline of an academic paper."""
 
-from typing import Dict, Self
+from typing import ClassVar, Dict, Type
 
-from fabricatio.fs.readers import extract_sections
 from fabricatio.models.extra.article_base import (
     ArticleBase,
     ChapterBase,
@@ -19,36 +18,14 @@ class ArticleSubsectionOutline(SubSectionBase):
 
 class ArticleSectionOutline(SectionBase[ArticleSubsectionOutline]):
     """A slightly more detailed research component specification for academic paper generation, Must contain subsections."""
-    @classmethod
-    def from_typst_code(cls, title: str, body: str, **kwargs) -> Self:
-        """Parse the given Typst code into an ArticleSectionOutline instance."""
-        return super().from_typst_code(
-            title,
-            body,
-            subsections=[
-                ArticleSubsectionOutline.from_typst_code(*pack)
-                for pack in extract_sections(body, level=3, section_char="=")
-            ],
-        )
 
+    child_type: ClassVar[Type[SubSectionBase]] = ArticleSubsectionOutline
 
 
 class ArticleChapterOutline(ChapterBase[ArticleSectionOutline]):
     """Macro-structural unit implementing standard academic paper organization. Must contain sections."""
 
-    @classmethod
-    def from_typst_code(cls, title: str, body: str, **kwargs) -> Self:
-        """Parse the given Typst code into an ArticleChapterOutline instance."""
-        return super().from_typst_code(
-            title,
-            body,
-            sections=[
-                ArticleSectionOutline.from_typst_code(*pack)
-                for pack in extract_sections(body, level=2, section_char="=")
-            ],
-
-        )
-
+    child_type: ClassVar[Type[SectionBase]] = ArticleSectionOutline
 
 
 class ArticleOutline(
@@ -58,21 +35,11 @@ class ArticleOutline(
 ):
     """Outline of an academic paper, containing chapters, sections, subsections."""
 
+    child_type: ClassVar[Type[ChapterBase]] = ArticleChapterOutline
+
     def _as_prompt_inner(self) -> Dict[str, str]:
         return {
             "Original Article Briefing": self.referenced.referenced,
             "Original Article Proposal": self.referenced.display(),
             "Original Article Outline": self.display(),
         }
-
-    @classmethod
-    def from_typst_code(cls, title: str, body: str, **kwargs) -> Self:
-        """Parse the given Typst code into an ArticleOutline instance."""
-        return super().from_typst_code(
-            title,
-            body,
-            chapters=[
-                ArticleChapterOutline.from_typst_code(*pack)
-                for pack in extract_sections(body, level=1, section_char="=")
-            ],
-        )

fabricatio/rust.pyi

@@ -16,6 +16,7 @@ from typing import Any, Dict, List, Optional, Tuple, overload
 
 from pydantic import JsonValue
 
+
 class TemplateManager:
     """Template rendering engine using Handlebars templates.
 
@@ -26,7 +27,7 @@ class TemplateManager:
     """
 
     def __init__(
-        self, template_dirs: List[Path], suffix: Optional[str] = None, active_loading: Optional[bool] = None
+            self, template_dirs: List[Path], suffix: Optional[str] = None, active_loading: Optional[bool] = None
     ) -> None:
         """Initialize the template manager.
 
@@ -58,8 +59,10 @@ class TemplateManager:
 
     @overload
     def render_template(self, name: str, data: Dict[str, Any]) -> str: ...
+
     @overload
     def render_template(self, name: str, data: List[Dict[str, Any]]) -> List[str]: ...
+
     def render_template(self, name: str, data: Dict[str, Any] | List[Dict[str, Any]]) -> str | List[str]:
         """Render a template with context data.
 
@@ -76,8 +79,10 @@ class TemplateManager:
 
     @overload
     def render_template_raw(self, template: str, data: Dict[str, Any]) -> str: ...
+
     @overload
     def render_template_raw(self, template: str, data: List[Dict[str, Any]]) -> List[str]: ...
+
     def render_template_raw(self, template: str, data: Dict[str, Any] | List[Dict[str, Any]]) -> str | List[str]:
         """Render a template with context data.
 
@@ -89,6 +94,7 @@ class TemplateManager:
             Rendered template content as string or list of strings
         """
 
+
 class BibManager:
     """BibTeX bibliography manager for parsing and querying citation data."""
 
@@ -197,6 +203,7 @@ class BibManager:
             Field value if found, None otherwise
         """
 
+
 def blake3_hash(content: bytes) -> str:
     """Calculate the BLAKE3 cryptographic hash of data.
 
@@ -207,9 +214,11 @@ def blake3_hash(content: bytes) -> str:
         Hex-encoded BLAKE3 hash string
     """
 
+
 def detect_language(string: str) -> str:
     """Detect the language of a given string."""
 
+
 def split_word_bounds(string: str) -> List[str]:
     """Split the string into words based on word boundaries.
 
@@ -220,6 +229,7 @@ def split_word_bounds(string: str) -> List[str]:
         A list of words extracted from the string.
     """
 
+
 def split_sentence_bounds(string: str) -> List[str]:
     """Split the string into sentences based on sentence boundaries.
 
@@ -230,6 +240,7 @@ def split_sentence_bounds(string: str) -> List[str]:
         A list of sentences extracted from the string.
     """
 
+
 def split_into_chunks(string: str, max_chunk_size: int, max_overlapping_rate: float = 0.3) -> List[str]:
     """Split the string into chunks of a specified size.
 
@@ -242,6 +253,7 @@ def split_into_chunks(string: str, max_chunk_size: int, max_overlapping_rate: fl
         A list of chunks extracted from the string.
     """
 
+
 def word_count(string: str) -> int:
     """Count the number of words in the string.
 
@@ -252,51 +264,67 @@ def word_count(string: str) -> int:
         The number of words in the string.
     """
 
+
 def is_chinese(string: str) -> bool:
     """Check if the given string is in Chinese."""
 
+
 def is_english(string: str) -> bool:
     """Check if the given string is in English."""
 
+
 def is_japanese(string: str) -> bool:
     """Check if the given string is in Japanese."""
 
+
 def is_korean(string: str) -> bool:
     """Check if the given string is in Korean."""
 
+
 def is_arabic(string: str) -> bool:
     """Check if the given string is in Arabic."""
 
+
 def is_russian(string: str) -> bool:
     """Check if the given string is in Russian."""
 
+
 def is_german(string: str) -> bool:
     """Check if the given string is in German."""
 
+
 def is_french(string: str) -> bool:
     """Check if the given string is in French."""
 
+
 def is_hindi(string: str) -> bool:
     """Check if the given string is in Hindi."""
 
+
 def is_italian(string: str) -> bool:
     """Check if the given string is in Italian."""
 
+
 def is_dutch(string: str) -> bool:
     """Check if the given string is in Dutch."""
 
+
 def is_portuguese(string: str) -> bool:
     """Check if the given string is in Portuguese."""
 
+
 def is_swedish(string: str) -> bool:
     """Check if the given string is in Swedish."""
 
+
 def is_turkish(string: str) -> bool:
     """Check if the given string is in Turkish."""
 
+
 def is_vietnamese(string: str) -> bool:
     """Check if the given string is in Vietnamese."""
 
+
 def tex_to_typst(string: str) -> str:
     """Convert TeX to Typst.
 
@@ -307,6 +335,7 @@ def tex_to_typst(string: str) -> str:
         The converted Typst string.
     """
 
+
 def convert_all_inline_tex(string: str) -> str:
     """Convert all inline TeX code in the string.
 
@@ -317,6 +346,7 @@ def convert_all_inline_tex(string: str) -> str:
         The converted string with inline TeX code replaced.
     """
 
+
 def convert_all_block_tex(string: str) -> str:
     """Convert all block TeX code in the string.
 
@@ -327,6 +357,7 @@ def convert_all_block_tex(string: str) -> str:
         The converted string with block TeX code replaced.
     """
 
+
 def fix_misplaced_labels(string: str) -> str:
     """A func to fix labels in a string.
 
@@ -337,6 +368,7 @@ def fix_misplaced_labels(string: str) -> str:
         The fixed string with labels properly placed.
     """
 
+
 def comment(string: str) -> str:
     """Add comment to the string.
 
@@ -347,6 +379,7 @@ def comment(string: str) -> str:
         The string with each line prefixed by '// '.
     """
 
+
 def uncomment(string: str) -> str:
     """Remove comment from the string.
 
@@ -357,6 +390,7 @@ def uncomment(string: str) -> str:
         The string with comments (lines starting with '// ' or '//') removed.
     """
 
+
 def split_out_metadata(string: str) -> Tuple[Optional[JsonValue], str]:
     """Split out metadata from a string.
 
@@ -367,6 +401,7 @@ def split_out_metadata(string: str) -> Tuple[Optional[JsonValue], str]:
         A tuple containing the metadata as a Python object (if parseable) and the remaining string.
     """
 
+
 def to_metadata(data: JsonValue) -> str:
     """Convert a Python object to a YAML string.
 
@@ -377,8 +412,37 @@ def to_metadata(data: JsonValue) -> str:
         The YAML string representation of the input data.
     """
 
+
 def convert_to_inline_formula(string: str) -> str:
     r"""Convert `$...$` to inline formula `\(...\)` and trim spaces."""
 
+
 def convert_to_block_formula(string: str) -> str:
     r"""Convert `$$...$$` to block formula `\[...\]` and trim spaces."""
+
+
+def inplace_update(string: str, wrapper: str, new_body: str) -> Optional[str]:
+    """Replace content between wrapper strings.
+
+    Args:
+        string: The input string containing content wrapped by delimiter strings.
+        wrapper: The delimiter string that marks the beginning and end of the content to replace.
+        new_body: The new content to place between the wrapper strings.
+
+    Returns:
+        A new string with the content between wrappers replaced.
+        
+    """
+
+
+def extract_body(string: str, wrapper: str) -> Optional[str]:
+    """
+    Extract the content between two occurrences of a wrapper string.
+
+    Args:
+        string: The input string containing content wrapped by delimiter strings.
+        wrapper: The delimiter string that marks the beginning and end of the content to extract.
+
+    Returns:
+        The content between the first two occurrences of the wrapper string if found, otherwise None.
+    """

fabricatio/utils.py

@@ -182,7 +182,7 @@ class RerankerAPI:
             response = requests.post(f"{self.base_url}/rerank", json=payload)
 
             # Handle non-200 status codes
-            if response.ok:
+            if not response.ok:
                 error_data = None
                 if "application/json" in response.headers.get("Content-Type", ""):
                     error_data = response.json()
@@ -239,7 +239,7 @@ class RerankerAPI:
                 session.post(f"{self.base_url}/rerank", json=payload) as response,
             ):
                 # Handle non-200 status codes
-                if response.ok:
+                if not response.ok:
                     if "application/json" in response.headers.get("Content-Type", ""):
                         error_data = await response.json()
                     else:

{fabricatio-0.2.13.dev1.dist-info → fabricatio-0.2.13.dev3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fabricatio
-Version: 0.2.13.dev1
+Version: 0.2.13.dev3
 Classifier: License :: OSI Approved :: MIT License
 Classifier: Programming Language :: Rust
 Classifier: Programming Language :: Python :: 3.12

{fabricatio-0.2.13.dev1.dist-info → fabricatio-0.2.13.dev3.dist-info}/RECORD

@@ -1,21 +1,21 @@
-fabricatio-0.2.13.dev1.dist-info/METADATA,sha256=ENDDnuCUK0GlyVyXGfL3-_wBsswudphZ1HOhiD-EN7Y,5316
-fabricatio-0.2.13.dev1.dist-info/WHEEL,sha256=jABKVkLC9kJr8mi_er5jOqpiQUjARSLXDUIIxDqsS50,96
-fabricatio-0.2.13.dev1.dist-info/licenses/LICENSE,sha256=do7J7EiCGbq0QPbMAL_FqLYufXpHnCnXBOuqVPwSV8Y,1088
+fabricatio-0.2.13.dev3.dist-info/METADATA,sha256=4VR-__6SDN1KEhH8cc6CMFqlD9dgq7Van2SValZiX1I,5316
+fabricatio-0.2.13.dev3.dist-info/WHEEL,sha256=jABKVkLC9kJr8mi_er5jOqpiQUjARSLXDUIIxDqsS50,96
+fabricatio-0.2.13.dev3.dist-info/licenses/LICENSE,sha256=do7J7EiCGbq0QPbMAL_FqLYufXpHnCnXBOuqVPwSV8Y,1088
 fabricatio/actions/article.py,sha256=c9HfLM1p0kkdEcQrl8w6Qa5p8riPnkEO--QbNZj88HQ,12562
-fabricatio/actions/article_rag.py,sha256=sG8eW75tDuICbGxBlMFzt0mPPNWqbMrR_Sugfv6jpyo,19100
+fabricatio/actions/article_rag.py,sha256=HJOiDblADZM1XXtBv3-v6lgh4gAFrS9LxKn5qKxR8Dw,18425
 fabricatio/actions/fs.py,sha256=gJR14U4ln35nt8Z7OWLVAZpqGaLnED-r1Yi-lX22tkI,959
 fabricatio/actions/output.py,sha256=lX0HkDse3ypCzZgeF-8Dr-EnNFdBiE-WQ1iLPFlGM1g,8302
 fabricatio/actions/rag.py,sha256=KN-OWgcQjGmNgSZ-s5B8m4LpYKSGFJR8eq72mo2CP9k,3592
 fabricatio/actions/rules.py,sha256=dkvCgNDjt2KSO1VgPRsxT4YBmIIMeetZb5tiz-slYkU,3640
 fabricatio/actions/__init__.py,sha256=wVENCFtpVb1rLFxoOFJt9-8smLWXuJV7IwA8P3EfFz4,48
 fabricatio/capabilities/advanced_judge.py,sha256=selB0Gwf1F4gGJlwBiRo6gI4KOUROgh3WnzO3mZFEls,706
-fabricatio/capabilities/advanced_rag.py,sha256=1LAAaCgVnZllNzLH3yW9MNwguUNPB7NMOJbD83VLdAo,2376
+fabricatio/capabilities/advanced_rag.py,sha256=FaXHGOqS4VleGSLsnC5qm4S4EBcHZLZbj8TjXRieKBs,2513
 fabricatio/capabilities/censor.py,sha256=bBT5qy-kp7fh8g4Lz3labSwxwJ60gGd_vrkc6k1cZ1U,4719
 fabricatio/capabilities/check.py,sha256=kYqzohhv2bZfl1aKSUt7a8snT8YEl2zgha_ZdAdMMfQ,8622
 fabricatio/capabilities/correct.py,sha256=W_cInqlciNEhyMK0YI53jk4EvW9uAdge90IO9OElUmA,10420
 fabricatio/capabilities/extract.py,sha256=PMjkWvbsv57IYT7zzd_xbIu4eQqQjpcmBtJzqlWZhHY,2495
 fabricatio/capabilities/propose.py,sha256=hkBeSlmcTdfYWT-ph6nlbtHXBozi_JXqXlWcnBy3W78,2007
-fabricatio/capabilities/rag.py,sha256=q2h4v3aXIKXspt7Xh6jBIf-NqPhBi2ispfrkmJEj13E,10743
+fabricatio/capabilities/rag.py,sha256=icJgFwEX4eOlZdtchDaITjFRBrAON6xezyFfPmzehI8,11058
 fabricatio/capabilities/rating.py,sha256=iMtQs3H6vCjuEjiuuz4SRKMVaX7yff7MHWz-slYvi5g,17835
 fabricatio/capabilities/review.py,sha256=-EMZe0ADFPT6fPGmra16UPjJC1M3rAs6dPFdTZ88Fgg,5060
 fabricatio/capabilities/task.py,sha256=uks1U-4LNCUdwdRxAbJJjMc31hOw6jlrcYriuQQfb04,4475
@@ -29,14 +29,14 @@ fabricatio/fs/readers.py,sha256=UXvcJO3UCsxHu9PPkg34Yh55Zi-miv61jD_wZQJgKRs,1751
 fabricatio/fs/__init__.py,sha256=FydmlEY_3QY74r1BpGDc5lFLhE6g6gkwOAtE30Fo-aI,786
 fabricatio/journal.py,sha256=stnEP88aUBA_GmU9gfTF2EZI8FS2OyMLGaMSTgK4QgA,476
 fabricatio/models/action.py,sha256=qxPeOD_nYNN94MzOhCzRDhySZFvM8uoZb_hhA7d_yn4,10609
-fabricatio/models/adv_kwargs_types.py,sha256=NfYEeY0WV3YVayqFTi54NC6WkGhLxCX4KNlsw_J2Nmk,2058
+fabricatio/models/adv_kwargs_types.py,sha256=IBV3ZcsNLvvEjO_2hBpYg_wLSpNKaMx6Ndam3qXJCw8,2097
 fabricatio/models/events.py,sha256=wiirk_ASg3iXDOZU_gIimci1VZVzWE1nDmxy-hQVJ9M,4150
 fabricatio/models/extra/advanced_judge.py,sha256=INUl_41C8jkausDekkjnEmTwNfLCJ23TwFjq2cM23Cw,1092
-fabricatio/models/extra/aricle_rag.py,sha256=p91JI8FmFSrHVg5KbhJq4w8vQdx4VUW75SrtQUi9ju4,10987
-fabricatio/models/extra/article_base.py,sha256=qBkRYAOdrtTnO02G0W1zDDtQWYrQIKot_XyyDaaCLp8,15697
+fabricatio/models/extra/aricle_rag.py,sha256=6mkuxNZD_8cdWINmxP8ajmTwdwSH45jcdUSBmY6ZZfQ,11685
+fabricatio/models/extra/article_base.py,sha256=0dVVbcv5GjjqJOGvobpM9CaIS876ymye8GH9EZ2jUCk,16762
 fabricatio/models/extra/article_essence.py,sha256=mlIkkRMR3I1RtqiiOnmIE3Vy623L4eECumkRzryE1pw,2749
-fabricatio/models/extra/article_main.py,sha256=0gvdodo0gpYSe9YjGhfXurT3oYzB5_xVqkT1oEDY9Z0,11148
-fabricatio/models/extra/article_outline.py,sha256=C9WNZNSz6gyuw9lDp29qidPvHBTENaeqcHfCoE_Y7F4,2793
+fabricatio/models/extra/article_main.py,sha256=zUkFQRLv6cLoPBbo7H21rrRScjgGv_SzhFd0Y514FsA,11211
+fabricatio/models/extra/article_outline.py,sha256=mw7eOuKMJgns4bihjcjOEIpAy38i0g-x6T6Vx3J0T5A,1629
 fabricatio/models/extra/article_proposal.py,sha256=NbyjW-7UiFPtnVD9nte75re4xL2pD4qL29PpNV4Cg_M,1870
 fabricatio/models/extra/patches.py,sha256=_WNCxtYzzsVfUxI16vu4IqsLahLYRHdbQN9er9tqhC0,997
 fabricatio/models/extra/problem.py,sha256=8tTU-3giFHOi5j7NJsvH__JJyYcaGrcfsRnkzQNm0Ew,7216
@@ -51,17 +51,17 @@ fabricatio/models/tool.py,sha256=jQ51g4lwTPfsMF1nbreDJtBczbxIHoXcPuLSOqHliq8,125
 fabricatio/models/usages.py,sha256=0bzITf0vug9ZaN6qnjNfFB7T8BAvpXE0bvx0otFYLLA,33356
 fabricatio/parser.py,sha256=-RbW2yzfJiu2ARq-lZw4tfgsjY2rIZWtJpoUmaE6gJQ,6637
 fabricatio/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-fabricatio/rust.pyi,sha256=HbdeWbVK7QL2Z3VHh2aMJ1sGUGZyYoOLzDtn5nQIv18,11591
+fabricatio/rust.pyi,sha256=ermwopmW2vbR05fQenayq4tbHIf6y0a64SMlp4ftoZA,12640
 fabricatio/rust_instances.py,sha256=Byeo8KHW_dJiXujJq7YPGDLBX5bHNDYbBc4sY3uubVY,313
 fabricatio/toolboxes/arithmetic.py,sha256=WLqhY-Pikv11Y_0SGajwZx3WhsLNpHKf9drzAqOf_nY,1369
 fabricatio/toolboxes/fs.py,sha256=l4L1CVxJmjw9Ld2XUpIlWfV0_Fu_2Og6d3E13I-S4aE,736
 fabricatio/toolboxes/__init__.py,sha256=KBJi5OG_pExscdlM7Bnt_UF43j4I3Lv6G71kPVu4KQU,395
-fabricatio/utils.py,sha256=oJ_PPhyRDPTnnmakIpBPXBm_y2PXHFcRVekFWxR_I8A,10291
+fabricatio/utils.py,sha256=IBKfs2Rg3bJnazzvj1-Fz1rMWNKhiuQG5_rZ1nxQeMI,10299
 fabricatio/workflows/articles.py,sha256=ObYTFUqLUk_CzdmmnX6S7APfxcGmPFqnFr9pdjU7Z4Y,969
 fabricatio/workflows/rag.py,sha256=-YYp2tlE9Vtfgpg6ROpu6QVO8j8yVSPa6yDzlN3qVxs,520
 fabricatio/workflows/__init__.py,sha256=5ScFSTA-bvhCesj3U9Mnmi6Law6N1fmh5UKyh58L3u8,51
 fabricatio/__init__.py,sha256=Rmvq2VgdS2u68vnOi2i5RbeWbAwrJDbk8D8D883PJWE,1022
-fabricatio/rust.cp312-win_amd64.pyd,sha256=jOC7VoxFTr8NSJyAR95446yUyqny3Bp6F2GlWxoxTYc,4449792
-fabricatio-0.2.13.dev1.data/scripts/tdown.exe,sha256=3bjfVSSSH6XXabKxMmS0tQLzoOnVQGTbDIZDnpJuEww,3353088
-fabricatio-0.2.13.dev1.data/scripts/ttm.exe,sha256=urASVLhELNfz0_gPdPU_LL-L9UCSRnluYOuJrmPwwt4,2556416
-fabricatio-0.2.13.dev1.dist-info/RECORD,,
+fabricatio/rust.cp312-win_amd64.pyd,sha256=hY2p1xhnfqN0T9y9S3oVhE3jjaRdISnpjx7GD07dg-w,4465664
+fabricatio-0.2.13.dev3.data/scripts/tdown.exe,sha256=xQabNheXRfFgZNVr_nOZkpbJ_O5RvJXn1nhtR45Vvaw,3359232
+fabricatio-0.2.13.dev3.data/scripts/ttm.exe,sha256=69ytRENft3g9IM_KG9NYulGvbeD4lZdp7QfcK2MJCVM,2554880
+fabricatio-0.2.13.dev3.dist-info/RECORD,,