epub-translator 0.1.6__py3-none-any.whl → 0.1.8__py3-none-any.whl

epub_translator/data/translate.jinja

@@ -13,6 +13,9 @@ Translation rules:
 {% if user_prompt -%}
 User may provide additional requirements in <rules> tags before the source text. Follow them, but prioritize the rules above if conflicts arise.
 
+<rules>
+{{ user_prompt }}
+</rules>
 {% endif -%}
 
 Output only the translated text, nothing else.

epub_translator/llm/core.py

@@ -13,6 +13,7 @@ from ..template import create_env
 from .context import LLMContext
 from .executor import LLMExecutor
 from .increasable import Increasable
+from .statistics import Statistics
 from .types import Message
 
 # Global state for logger filename generation
@@ -44,7 +45,7 @@ class LLM:
         self._temperature: Increasable = Increasable(temperature)
         self._cache_path: Path | None = self._ensure_dir_path(cache_path)
         self._logger_save_path: Path | None = self._ensure_dir_path(log_dir_path)
-
+        self._statistics = Statistics()
         self._executor = LLMExecutor(
             url=url,
             model=model,
@@ -53,12 +54,29 @@ class LLM:
             retry_times=retry_times,
             retry_interval_seconds=retry_interval_seconds,
             create_logger=self._create_logger,
+            statistics=self._statistics,
         )
 
     @property
     def encoding(self) -> Encoding:
         return self._encoding
 
+    @property
+    def total_tokens(self) -> int:
+        return self._statistics.total_tokens
+
+    @property
+    def input_tokens(self) -> int:
+        return self._statistics.input_tokens
+
+    @property
+    def input_cache_tokens(self) -> int:
+        return self._statistics.input_cache_tokens
+
+    @property
+    def output_tokens(self) -> int:
+        return self._statistics.output_tokens
+
     def context(self, cache_seed_content: str | None = None) -> LLMContext:
         return LLMContext(
             executor=self._executor,

epub_translator/llm/executor.py

@@ -7,6 +7,7 @@ from openai import OpenAI
 from openai.types.chat import ChatCompletionMessageParam
 
 from .error import is_retry_error
+from .statistics import Statistics
 from .types import Message, MessageRole
 
 
@@ -20,12 +21,14 @@ class LLMExecutor:
         retry_times: int,
         retry_interval_seconds: float,
         create_logger: Callable[[], Logger | None],
+        statistics: Statistics,
     ) -> None:
         self._model_name: str = model
         self._timeout: float | None = timeout
         self._retry_times: int = retry_times
         self._retry_interval_seconds: float = retry_interval_seconds
         self._create_logger: Callable[[], Logger | None] = create_logger
+        self._statistics = statistics
         self._client = OpenAI(
             api_key=api_key,
             base_url=url,
@@ -156,6 +159,7 @@ class LLMExecutor:
             model=self._model_name,
             messages=messages,
             stream=True,
+            stream_options={"include_usage": True},
             top_p=top_p,
             temperature=temperature,
             max_tokens=max_tokens,
@@ -164,4 +168,5 @@ class LLMExecutor:
         for chunk in stream:
             if chunk.choices and chunk.choices[0].delta.content:
                 buffer.write(chunk.choices[0].delta.content)
+            self._statistics.submit_usage(chunk.usage)
         return buffer.getvalue()

epub_translator/llm/statistics.py

@@ -0,0 +1,25 @@
+from threading import Lock
+
+from openai.types import CompletionUsage
+
+
+class Statistics:
+    def __init__(self) -> None:
+        self._lock = Lock()
+        self.total_tokens = 0
+        self.input_tokens = 0
+        self.input_cache_tokens = 0
+        self.output_tokens = 0
+
+    def submit_usage(self, usage: CompletionUsage | None) -> None:
+        if usage is None:
+            return
+        with self._lock:
+            if usage.total_tokens:
+                self.total_tokens += usage.total_tokens
+            if usage.prompt_tokens:
+                self.input_tokens += usage.prompt_tokens
+            if usage.prompt_tokens_details and usage.prompt_tokens_details.cached_tokens:
+                self.input_cache_tokens += usage.prompt_tokens_details.cached_tokens
+            if usage.completion_tokens:
+                self.output_tokens += usage.completion_tokens

epub_translator/segment/__init__.py

@@ -21,6 +21,7 @@ from .text_segment import (
     TextPosition,
     TextSegment,
     combine_text_segments,
+    find_block_depth,
     incision_between,
     search_text_segments,
 )

epub_translator/segment/text_segment.py

@@ -4,7 +4,12 @@ from enum import Enum, auto
 from typing import Self
 from xml.etree.ElementTree import Element
 
-from ..xml import expand_left_element_texts, expand_right_element_texts, is_inline_tag, normalize_text_in_element
+from ..xml import (
+    expand_left_element_texts,
+    expand_right_element_texts,
+    is_inline_element,
+    normalize_text_in_element,
+)
 
 
 class TextPosition(Enum):
@@ -100,7 +105,7 @@ def search_text_segments(root: Element) -> Generator[TextSegment, None, None]:
 def _search_text_segments(stack: list[Element], element: Element) -> Generator[TextSegment, None, None]:
     text = normalize_text_in_element(element.text)
     next_stack = stack + [element]
-    next_block_depth = _find_block_depth(next_stack)
+    next_block_depth = find_block_depth(next_stack)
 
     if text is not None:
         yield TextSegment(
@@ -125,12 +130,11 @@ def _search_text_segments(stack: list[Element], element: Element) -> Generator[T
             )
 
 
-def _find_block_depth(parent_stack: list[Element]) -> int:
+def find_block_depth(parent_stack: list[Element]) -> int:
     index: int = 0
-    for i in range(len(parent_stack) - 1, -1, -1):
-        if not is_inline_tag(parent_stack[i].tag):
+    for i in range(len(parent_stack)):
+        if not is_inline_element(parent_stack[i]):
             index = i
-            break
     return index + 1  # depth is a count not index
 
 

epub_translator/segment/utils.py

@@ -8,22 +8,6 @@ def element_fingerprint(element: Element) -> str:
     return f"<{element.tag} {' '.join(attrs)}/>"
 
 
-def unwrap_parents(element: Element) -> tuple[Element, list[Element]]:
-    parents: list[Element] = []
-    while True:
-        if len(element) != 1:
-            break
-        child = element[0]
-        if not element.text:
-            break
-        if not child.tail:
-            break
-        parents.append(element)
-        element = child
-        element.tail = None
-    return element, parents
-
-
 def id_in_element(element: Element) -> int | None:
     id_str = element.get(ID_KEY, None)
     if id_str is None:

epub_translator/translation/xml_interrupter.py

@@ -1,9 +1,13 @@
 from collections.abc import Generator, Iterable
 from typing import cast
-from xml.etree.ElementTree import Element
+from xml.etree.ElementTree import Element, tostring
 
-from ..segment import TextSegment
+from bs4 import BeautifulSoup
+from mathml2latex.mathml import process_mathml
+
+from ..segment import TextSegment, combine_text_segments, find_block_depth
 from ..utils import ensure_list, normalize_whitespace
+from ..xml import DISPLAY_ATTRIBUTE, clone_element, is_inline_element
 
 _ID_KEY = "__XML_INTERRUPTER_ID"
 _MATH_TAG = "math"
@@ -37,8 +41,10 @@ class XMLInterrupter:
     def interrupt_block_element(self, element: Element) -> Element:
         interrupted_element = self._placeholder2interrupted.pop(id(element), None)
         if interrupted_element is None:
+            element.attrib.pop(_ID_KEY, None)
             return element
         else:
+            interrupted_element.attrib.pop(_ID_KEY, None)
             return interrupted_element
 
     def _expand_source_text_segment(self, text_segment: TextSegment):
@@ -81,14 +87,18 @@ class XMLInterrupter:
                     _ID_KEY: cast(str, interrupted_element.get(_ID_KEY)),
                 },
             )
+            interrupted_display = interrupted_element.get(DISPLAY_ATTRIBUTE, None)
+            if interrupted_display is not None:
+                placeholder_element.set(DISPLAY_ATTRIBUTE, interrupted_display)
+
             raw_parent_stack = text_segment.parent_stack[:interrupted_index]
             parent_stack = raw_parent_stack + [placeholder_element]
             merged_text_segment = TextSegment(
-                text="".join(t.text for t in text_segments),
+                text=self._render_latex(text_segments),
                 parent_stack=parent_stack,
                 left_common_depth=text_segments[0].left_common_depth,
                 right_common_depth=text_segments[-1].right_common_depth,
-                block_depth=len(parent_stack),
+                block_depth=find_block_depth(parent_stack),
                 position=text_segments[0].position,
             )
             self._placeholder2interrupted[id(placeholder_element)] = interrupted_element
@@ -116,8 +126,8 @@ class XMLInterrupter:
                 # 原始栈退光，仅留下相对 interrupted 元素的栈，这种格式与 translated 要求一致
                 text_segment.left_common_depth = max(0, text_segment.left_common_depth - interrupted_index)
                 text_segment.right_common_depth = max(0, text_segment.right_common_depth - interrupted_index)
-                text_segment.block_depth = 1
                 text_segment.parent_stack = text_segment.parent_stack[interrupted_index:]
+                text_segment.block_depth = find_block_depth(text_segment.parent_stack)
 
         return merged_text_segment
 
@@ -129,37 +139,54 @@ class XMLInterrupter:
                 break
         return interrupted_index
 
+    def _render_latex(self, text_segments: list[TextSegment]) -> str:
+        math_element, _ = next(combine_text_segments(text_segments))
+        while math_element.tag != _MATH_TAG:
+            if len(math_element) == 0:
+                return ""
+            math_element = math_element[0]
+
+        math_element = clone_element(math_element)
+        math_element.attrib.pop(_ID_KEY, None)
+        math_element.tail = None
+        latex: str | None = None
+        try:
+            mathml_str = tostring(math_element, encoding="unicode")
+            soup = BeautifulSoup(mathml_str, "html.parser")
+            latex = process_mathml(soup)
+        except Exception:
+            pass
+
+        if latex is None:
+            latex = "".join(t.text for t in text_segments)
+            latex = normalize_whitespace(latex).strip()
+        else:
+            latex = normalize_whitespace(latex).strip()
+            if is_inline_element(math_element):
+                latex = f"${latex}$"
+            else:
+                latex = f"$${latex}$$"
+
+        return f" {latex} "
+
     def _expand_translated_text_segment(self, text_segment: TextSegment):
-        interrupted_id = text_segment.block_parent.attrib.pop(_ID_KEY, None)
+        parent_element = text_segment.parent_stack[-1]
+        interrupted_id = parent_element.attrib.pop(_ID_KEY, None)
         if interrupted_id is None:
             yield text_segment
             return
 
-        raw_text_segments = self._raw_text_segments.pop(interrupted_id, None)
-        if not raw_text_segments:
+        if parent_element is text_segment.block_parent:
+            # Block-level math， need to be hidden
             return
 
-        raw_block = raw_text_segments[0].parent_stack[0]
-        if not self._is_inline_math(raw_block):
+        raw_text_segments = self._raw_text_segments.pop(interrupted_id, None)
+        if not raw_text_segments:
+            yield text_segment
             return
 
         for raw_text_segment in raw_text_segments:
+            text_basic_parent_stack = text_segment.parent_stack[:-1]
             raw_text_segment.block_parent.attrib.pop(_ID_KEY, None)
+            raw_text_segment.parent_stack = text_basic_parent_stack + raw_text_segment.parent_stack
             yield raw_text_segment
-
-    def _has_no_math_texts(self, element: Element):
-        if element.tag == _MATH_TAG:
-            return True
-        if element.text and normalize_whitespace(element.text).strip():
-            return False
-        for child_element in element:
-            if not self._has_no_math_texts(child_element):
-                return False
-            if child_element.tail and normalize_whitespace(child_element.tail).strip():
-                return False
-        return True
-
-    def _is_inline_math(self, element: Element) -> bool:
-        if element.tag != _MATH_TAG:
-            return False
-        return element.get("display", "").lower() != "block"

epub_translator/xml/const.py

@@ -1 +1,2 @@
 ID_KEY: str = "id"
+DISPLAY_ATTRIBUTE: str = "display"

epub_translator/xml/inline.py

@@ -1,6 +1,11 @@
+from xml.etree.ElementTree import Element
+
+from .const import DISPLAY_ATTRIBUTE
+
 # HTML inline-level elements
 # Reference: https://developer.mozilla.org/en-US/docs/Web/HTML/Inline_elements
 # Reference: https://developer.mozilla.org/en-US/docs/Glossary/Inline-level_content
+# Reference: https://developer.mozilla.org/en-US/docs/MathML/Element
 _HTML_INLINE_TAGS = frozenset(
     (
         # Inline text semantics
@@ -59,9 +64,57 @@ _HTML_INLINE_TAGS = frozenset(
         "del",
         "ins",
         "slot",
+        # MathML elements
+        # Token elements
+        "mi",  # identifier
+        "mn",  # number
+        "mo",  # operator
+        "ms",  # string literal
+        "mspace",  # space
+        "mtext",  # text
+        # General layout
+        "menclose",  # enclosed content
+        "merror",  # syntax error message
+        "mfenced",  # parentheses (deprecated)
+        "mfrac",  # fraction
+        "mpadded",  # space around content
+        "mphantom",  # invisible content
+        "mroot",  # radical with index
+        "mrow",  # grouped sub-expressions
+        "msqrt",  # square root
+        "mstyle",  # style change
+        # Scripts and limits
+        "mmultiscripts",  # prescripts and tensor indices
+        "mover",  # overscript
+        "mprescripts",  # prescripts separator
+        "msub",  # subscript
+        "msubsup",  # subscript-superscript pair
+        "msup",  # superscript
+        "munder",  # underscript
+        "munderover",  # underscript-overscript pair
+        # Table math
+        "mtable",  # table or matrix
+        "mtr",  # row in table or matrix
+        "mtd",  # cell in table or matrix
+        # Semantic annotations
+        "annotation",  # data annotation
+        "annotation-xml",  # XML annotation
+        "semantics",  # semantic annotation container
+        # Other
+        "maction",  # bind actions to sub-expressions (deprecated)
     )
 )
 
 
-def is_inline_tag(tag: str) -> bool:
-    return tag.lower() in _HTML_INLINE_TAGS
+def is_inline_element(element: Element) -> bool:
+    tag = element.tag.lower()
+    if tag in _HTML_INLINE_TAGS:
+        return True
+    display = element.get(DISPLAY_ATTRIBUTE, None)
+    if display is not None:
+        display = display.lower()
+        if display == "inline":
+            return True
+    if tag == "math" and display != "block":
+        return True
+    return False

epub_translator/xml_translator/submitter.py

@@ -4,7 +4,7 @@ from enum import Enum, auto
 from xml.etree.ElementTree import Element
 
 from ..segment import TextSegment, combine_text_segments
-from ..xml import index_of_parent, is_inline_tag, iter_with_stack
+from ..xml import index_of_parent, is_inline_element, iter_with_stack
 from .stream_mapper import InlineSegmentMapping
 
 
@@ -78,7 +78,7 @@ class _Submitter:
         preserved_elements: list[Element] = []
         if self._action == SubmitKind.REPLACE:
             for child in list(node.raw_element):
-                if not is_inline_tag(child.tag):
+                if not is_inline_element(child):
                     child.tail = None
                     preserved_elements.append(child)
 
@@ -87,7 +87,7 @@ class _Submitter:
 
         if combined is not None:
             # 在 APPEND_BLOCK 模式下，如果是 inline tag，则在文本前面加空格
-            if self._action == SubmitKind.APPEND_BLOCK and is_inline_tag(combined.tag) and combined.text:
+            if self._action == SubmitKind.APPEND_BLOCK and is_inline_element(combined) and combined.text:
                 combined.text = " " + combined.text
             parent.insert(index + 1, combined)
             index += 1
@@ -200,7 +200,7 @@ class _Submitter:
         preserved_elements: list[Element] = []
         for i in range(start_index, end_index):
             elem = node_element[i]
-            if not is_inline_tag(elem.tag):
+            if not is_inline_element(elem):
                 elem.tail = None
                 preserved_elements.append(elem)
 
@@ -223,7 +223,7 @@ class _Submitter:
 
         if combined.text:
             will_inject_space = self._action == SubmitKind.APPEND_TEXT or (
-                is_inline_tag(combined.tag) and self._action == SubmitKind.APPEND_BLOCK
+                is_inline_element(combined) and self._action == SubmitKind.APPEND_BLOCK
             )
             if tail_element is not None:
                 tail_element.tail = self._append_text_in_element(

{epub_translator-0.1.6.dist-info → epub_translator-0.1.8.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: epub-translator
-Version: 0.1.6
+Version: 0.1.8
 Summary: Translate the epub book using LLM. The translated book will retain the original text and list the translated text side by side with the original text.
 License: MIT
 Keywords: epub,llm,translation,translator
@@ -24,6 +24,7 @@ Classifier: Topic :: Software Development :: Localization
 Classifier: Topic :: Text Processing :: Markup
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Requires-Dist: jinja2 (>=3.1.6,<4.0.0)
+Requires-Dist: mathml2latex (>=0.2.12,<0.3.0)
 Requires-Dist: openai (>=2.14.0,<3.0.0)
 Requires-Dist: resource-segmentation (>=0.0.7,<0.1.0)
 Requires-Dist: tiktoken (>=0.12.0,<1.0.0)
@@ -59,6 +60,13 @@ Translate EPUB books using Large Language Models while preserving the original t
 - **Flexible LLM Support**: Works with any OpenAI-compatible API endpoint
 - **Caching**: Built-in caching for progress recovery when translation fails
 
+## Use Cases
+
+- **Language Learning**: Read books in their original language with side-by-side translations
+- **Academic Research**: Access foreign literature with bilingual references
+- **Content Localization**: Prepare books for international audiences
+- **Cross-Cultural Reading**: Enjoy literature while understanding cultural nuances
+
 ## Installation
 
 ```bash
@@ -357,13 +365,6 @@ llm = LLM(
 )
 ```
 
-## Use Cases
-
-- **Language Learning**: Read books in their original language with side-by-side translations
-- **Academic Research**: Access foreign literature with bilingual references
-- **Content Localization**: Prepare books for international audiences
-- **Cross-Cultural Reading**: Enjoy literature while understanding cultural nuances
-
 ## Advanced Features
 
 ### Custom Translation Prompts
@@ -421,6 +422,105 @@ translate(
 
 When using `concurrency > 1`, ensure that any custom callback functions (`on_progress`, `on_fill_failed`) are thread-safe. Built-in callbacks are thread-safe by default.
 
+### Token Usage Monitoring
+
+Track token consumption during translation to monitor API costs and usage:
+
+```python
+from epub_translator import LLM, translate, language, SubmitKind
+
+llm = LLM(
+    key="your-api-key",
+    url="https://api.openai.com/v1",
+    model="gpt-4",
+    token_encoding="o200k_base",
+)
+
+translate(
+    source_path="source.epub",
+    target_path="translated.epub",
+    target_language=language.ENGLISH,
+    submit=SubmitKind.APPEND_BLOCK,
+    llm=llm,
+)
+
+# Access token statistics after translation
+print(f"Total tokens: {llm.total_tokens}")
+print(f"Input tokens: {llm.input_tokens}")
+print(f"Input cache tokens: {llm.input_cache_tokens}")
+print(f"Output tokens: {llm.output_tokens}")
+```
+
+**Available Statistics:**
+
+- `total_tokens` - Total number of tokens used (input + output)
+- `input_tokens` - Number of prompt/input tokens
+- `input_cache_tokens` - Number of cached input tokens (when using prompt caching)
+- `output_tokens` - Number of generated/completion tokens
+
+**Real-time Monitoring:**
+
+You can also monitor token usage in real-time during translation:
+
+```python
+from tqdm import tqdm
+import time
+
+with tqdm(total=100, desc="Translating", unit="%") as pbar:
+    last_progress = 0.0
+    start_time = time.time()
+
+    def on_progress(progress: float):
+        nonlocal last_progress
+        increment = (progress - last_progress) * 100
+        pbar.update(increment)
+        last_progress = progress
+
+        # Update token stats in progress bar
+        pbar.set_postfix({
+            'tokens': llm.total_tokens,
+            'cost_est': f'${llm.total_tokens * 0.00001:.4f}'  # Estimate based on your pricing
+        })
+
+    translate(
+        source_path="source.epub",
+        target_path="translated.epub",
+        target_language=language.ENGLISH,
+        submit=SubmitKind.APPEND_BLOCK,
+        llm=llm,
+        on_progress=on_progress,
+    )
+
+    elapsed = time.time() - start_time
+    print(f"\nTranslation completed in {elapsed:.1f}s")
+    print(f"Total tokens used: {llm.total_tokens:,}")
+    print(f"Average tokens/second: {llm.total_tokens/elapsed:.1f}")
+```
+
+**Dual-LLM Token Tracking:**
+
+When using separate LLMs for translation and filling, each LLM tracks its own statistics:
+
+```python
+translation_llm = LLM(key="...", url="...", model="gpt-4", token_encoding="o200k_base")
+fill_llm = LLM(key="...", url="...", model="gpt-4", token_encoding="o200k_base")
+
+translate(
+    source_path="source.epub",
+    target_path="translated.epub",
+    target_language=language.ENGLISH,
+    submit=SubmitKind.APPEND_BLOCK,
+    translation_llm=translation_llm,
+    fill_llm=fill_llm,
+)
+
+print(f"Translation tokens: {translation_llm.total_tokens}")
+print(f"Fill tokens: {fill_llm.total_tokens}")
+print(f"Combined total: {translation_llm.total_tokens + fill_llm.total_tokens}")
+```
+
+**Note:** Token statistics are cumulative across all API calls made by the LLM instance. The counts only increase and are thread-safe when using concurrent translation.
+
 ## Related Projects
 
 ### PDF Craft

{epub_translator-0.1.6.dist-info → epub_translator-0.1.8.dist-info}/RECORD

@@ -1,14 +1,6 @@
 epub_translator/__init__.py,sha256=JsiOUPpk5k7q8mXIgnRQWdVVnkJww_KDTg7jXsP7_C4,222
 epub_translator/data/fill.jinja,sha256=zSytA8Vhp2i6YBZ09F1z9iPJq1-jUaiphoXqTNZwnvo,6964
-epub_translator/data/mmltex/README.md,sha256=wwhe5yW1U_7_YZIFKnQVnCOmUl7Mu3gsr3lNnDSJ5Qs,2953
-epub_translator/data/mmltex/cmarkup.xsl,sha256=DkhimAATM0XSCfVOfY41-qTPoddqzOHjZ00Pynr4zQE,37707
-epub_translator/data/mmltex/entities.xsl,sha256=TYZ5iGg0u9XlDDBBGuZiHL7MsxKc-3OsTIBAVM1GDek,107742
-epub_translator/data/mmltex/glayout.xsl,sha256=Ztc7N1wiHaYZlo9u9iuROrIl3uIIIoo1VFIuojXq7TM,6879
-epub_translator/data/mmltex/mmltex.xsl,sha256=BVXFbApz-9W2qRKKtBTxptK5vxG2bfB8tv9W1MP5iBI,1384
-epub_translator/data/mmltex/scripts.xsl,sha256=f4ei0cDCW3cV-Ra7rC3kC5tRcKdjJxbSpCeQLoohtgo,13697
-epub_translator/data/mmltex/tables.xsl,sha256=RxtNo8qDtVAg8_6BuYsafraB_0z7YDAB9D__fT9gmWs,4327
-epub_translator/data/mmltex/tokens.xsl,sha256=j3JZRcBhAiiY8o5K3640phfLwxO8JVspCFlSttwBzJk,12373
-epub_translator/data/translate.jinja,sha256=93d8kschm5HV-EfXd1kFSIVMObDqTMdoUrwDfce2bhU,820
+epub_translator/data/translate.jinja,sha256=MVAWvgO9kybEFi0zMiZLEWwuRUL3l8PrwJdsoueQeCs,855
 epub_translator/epub/__init__.py,sha256=aZawPakdkEquL4kRRpyCTdoSQ82l7FGqY4Uw6-ndoGA,154
 epub_translator/epub/common.py,sha256=4-SpTe8iot9hMfyXILmlUFvYVNYqPAHL5hn1fr2wgis,1180
 epub_translator/epub/math.py,sha256=-Q2LJQxxjgQZQUe_WlJA9tjzLqgqtw2ZmbGbHsPRp2U,5422
@@ -18,17 +10,18 @@ epub_translator/epub/toc.py,sha256=TKJfyDT4svFkXd6JCNZk2ZEYc9q-5DXnV3zY2UKo8nE,1
 epub_translator/epub/zip.py,sha256=-3LI8f-ksgU8xCy28NjBOKyQPE8PhPEUPqIKZE1p8dw,2364
 epub_translator/llm/__init__.py,sha256=YcFYYnXmXyX0RUyC-PDbj5k7Woygp_XOpTI3vDiNSPM,75
 epub_translator/llm/context.py,sha256=8-0UnrZIaNshR_imy_ed_UpOK7H1a6dOsG-boaYOX8k,4186
-epub_translator/llm/core.py,sha256=wQwt6oG68ZN_iQOaytBiPXOC7sI62XII_A4dOHdAt_s,5979
+epub_translator/llm/core.py,sha256=MnToX8Zhr_r4sj9B3s54bclesojQEFarzl0VqHGDKlo,6488
 epub_translator/llm/error.py,sha256=4efAIQL14DFSvAnSTUfgdAbZRqaWBqOfUGsSfvxa5zM,1503
-epub_translator/llm/executor.py,sha256=A0IjQ-s9wBJuhAZAAydneb9zBXWnu2J9inR2Q8F-GDE,5533
+epub_translator/llm/executor.py,sha256=wxgFwWaLmuqAvctO3lcQX4U52aiw7EdaFw9Ut0v-ZzU,5745
 epub_translator/llm/increasable.py,sha256=8XkKeI1hiHlpMHj8dQ4fW0BkViSx4hH8QfbQsy-5SDw,1297
+epub_translator/llm/statistics.py,sha256=BX75qVWJ9aWbMoFtaQzoE8oVCLh7wiHoR06dX-AAl3E,875
 epub_translator/llm/types.py,sha256=c-dMAIvlG4R3la3mUTWEw5xei-sIYKmQeBja7mirxcI,219
-epub_translator/segment/__init__.py,sha256=UYTv_IKQbEB0DzhFeiuqCvjoJLvB-7XRwlaFS90KmIw,573
+epub_translator/segment/__init__.py,sha256=nCHNaHASElKTbC8HEAQkI1Y12m6kEdX5uJVvVvHKtFg,595
 epub_translator/segment/block_segment.py,sha256=psNKA_HMIcwZtoug8AtnAcV9_mQ2WXLnXqFsekHzt2g,4570
 epub_translator/segment/common.py,sha256=gGWYQaJ0tGnWCuF1me9TOo-Q_DrZVakCu2patyFIOs0,714
 epub_translator/segment/inline_segment.py,sha256=nrRKoJ-vblsNITJeixrCgIOkVQyUXrchMg0XYU_8pLo,14563
-epub_translator/segment/text_segment.py,sha256=LhGlugp6MeAB3tk2jxd1kBb2EA8G2ruN49mP_IZehA0,6295
-epub_translator/segment/utils.py,sha256=qMqUt33pDRN5Tnuydkodzu2gaQrwTzAnQmXpDuHen1o,1036
+epub_translator/segment/text_segment.py,sha256=E_qgPI09sCV_-PsJtgwcloTa0tpOP3wl0pw5gV9dDNY,6288
+epub_translator/segment/utils.py,sha256=_tlIA1I7rYz9_q-oQ5cPZWPmhTObCXjksQzRtX3beXY,636
 epub_translator/serial/__init__.py,sha256=b3IMVmWcUwEqHKcGmey88b057pyz5ct946CaUZi4LB4,67
 epub_translator/serial/chunk.py,sha256=FrTaHikVOd6bLYumnEriTaAQ_DIDLjHm16gh-wBVR9k,1495
 epub_translator/serial/segment.py,sha256=uEz-ke1KcYrON-68FaUEzMG2CzHlMjvbC11F3ZT4yH0,446
@@ -39,10 +32,10 @@ epub_translator/translation/epub_transcode.py,sha256=_pRzmQgDrlfsibalkUogVi0F0Qy
 epub_translator/translation/language.py,sha256=88osG0JNYxOkxBjg5Pm-P0Mhiyxf6GqdxoPW12HW0PE,493
 epub_translator/translation/punctuation.py,sha256=TPCGjEmlAyN3G11VuXdHn-pvUkuWDwWqbTNzw-ij60E,813
 epub_translator/translation/translator.py,sha256=WC4Yqx-ffhxBhqzMAujE_NQG7BsDwgn95UMNG7OkUSo,6487
-epub_translator/translation/xml_interrupter.py,sha256=QxrNpBoR4ZIAvWsa20jz1z_bE_5-G5-nBGjE6IKCTjw,7405
+epub_translator/translation/xml_interrupter.py,sha256=7TRGskn_OxRZT5mvKfjL0VMtU2VCgl1d9ElmfhFG0pM,8628
 epub_translator/utils.py,sha256=BfZWrYjzDNQ4cFrgvRNzd4i1CKLtPxS8Z4LBHhqEV78,914
 epub_translator/xml/__init__.py,sha256=qluFTfZYlPmOie8nR2C5O0tZ3UbCQEoEoR-Fq-__79c,160
-epub_translator/xml/const.py,sha256=Re2TYmpwG7-jVVgSq3R_K-uYhvAYzcXcRmLFkwCPD9Y,19
+epub_translator/xml/const.py,sha256=tCdeJfGwH5xgS4uOmR-pXSfyWXGxOHMJyZKE46BVkJU,54
 epub_translator/xml/deduplication.py,sha256=TaMbzeA70VvUQV0X1wcQFVbuMEPJUtj9Hq6iWlUmtAQ,1152
 epub_translator/xml/friendly/__init__.py,sha256=I5jhnhFWoHvojLsYXH4jfR4Gi8lKFZ3yQ56ze5hEe1M,74
 epub_translator/xml/friendly/decoder.py,sha256=xRQ5LnSunmYbba_0oT39oUr86-sLYAHYMUGmlseIu2U,2467
@@ -50,7 +43,7 @@ epub_translator/xml/friendly/encoder.py,sha256=evjvw6oE-oCud44IsJ-YZVHn6dtUzjNYX
 epub_translator/xml/friendly/parser.py,sha256=QlMHA0nfPJbNyx6IwRFrYVw7okuvzDB42NXCauIFV-o,6560
 epub_translator/xml/friendly/tag.py,sha256=ahaGoYttuAlnFxLFFgTV51KUZSpUiHho-COZX14nxN8,3308
 epub_translator/xml/friendly/transform.py,sha256=5tG1MJmzrXIR_Z5gmRxwcoKvXBzJBVH0ELeaRsG-8w0,1201
-epub_translator/xml/inline.py,sha256=mwFho6wq2gYWmWcg5Cw6OQeteV-a-i6X9OE63fzblpE,1274
+epub_translator/xml/inline.py,sha256=VcaNEF2ebVl2fogVk2yV3f4vOP4rePsPTV_qU3fJCE0,3108
 epub_translator/xml/self_closing.py,sha256=41ofGUdss9yU51IVwI4It6hKfzh8YcxIR_j-ohD19LE,5240
 epub_translator/xml/utils.py,sha256=7tQ6L5P0_JXhxONeG64hEeeL5mKjA6NKS1H1Q9B1Cac,1062
 epub_translator/xml/xml.py,sha256=qQ5Wk1-KVVHE4TX25zGOR7fINsGkXnoq-qyKKNl5no4,1675
@@ -62,10 +55,10 @@ epub_translator/xml_translator/concurrency.py,sha256=ACwoDHNX3xChL0On5yvUSFT8By7
 epub_translator/xml_translator/hill_climbing.py,sha256=1jvilOkTLzwljJA4Nrel8yU2XGvOXpueUJTK7RAp-XY,4272
 epub_translator/xml_translator/score.py,sha256=TkXDmr-29p8SzuAp68u_vFDE69y1TyId9S20HT1T_xs,5311
 epub_translator/xml_translator/stream_mapper.py,sha256=nk8iRUHAUQA2B35_y-JOCo6il8MSxXikWvyl-WA8WAA,10662
-epub_translator/xml_translator/submitter.py,sha256=6PGQTnEcOgL3zseDpSzDmU5d9Eg3eO5OfPIGmQp2DVY,14155
+epub_translator/xml_translator/submitter.py,sha256=_ic2_JBPdEd2nMSu2mtQ5OzqpGv0zGrvYaicVUXAiUQ,14159
 epub_translator/xml_translator/translator.py,sha256=7Ja1jFbmjIgHcmI9V6gg_K0t7qb6in9mhRn54a7qhZ8,9497
 epub_translator/xml_translator/validation.py,sha256=-OKlSZuD__sjAiEpGAO93YQme4ZDSPmoPjRsAMOCEjc,16668
-epub_translator-0.1.6.dist-info/LICENSE,sha256=5RF32sL3LtMOJIErdDKp1ZEYPGXS8WPpsiSz_jMBnGI,1066
-epub_translator-0.1.6.dist-info/METADATA,sha256=AcjUb1wmz6cN8PnbwgWJeGlOO9sH445B-qPugLW705M,15638
-epub_translator-0.1.6.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
-epub_translator-0.1.6.dist-info/RECORD,,
+epub_translator-0.1.8.dist-info/LICENSE,sha256=5RF32sL3LtMOJIErdDKp1ZEYPGXS8WPpsiSz_jMBnGI,1066
+epub_translator-0.1.8.dist-info/METADATA,sha256=DTipkbLL2pnijg7XIXSHogZXJzI009K7ZTkGUMy06d8,18663
+epub_translator-0.1.8.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
+epub_translator-0.1.8.dist-info/RECORD,,