epstein-files 1.0.0__py3-none-any.whl → 1.0.1__py3-none-any.whl

epstein_files/util/logging.py

@@ -0,0 +1,57 @@
+import logging
+from os import environ
+
+from rich.console import Console
+from rich.highlighter import ReprHighlighter
+from rich.logging import RichHandler
+from rich.theme import Theme
+
+from epstein_files.util.constant.strings import *
+
+FILENAME_STYLE = 'gray27'
+
+DOC_TYPE_STYLES = {
+    DOCUMENT_CLASS: 'grey69',
+    EMAIL_CLASS: 'dark_orange3',
+    JSON_FILE_CLASS: 'sandy_brown',
+    MESSENGER_LOG_CLASS: 'deep_pink4',
+    OTHER_FILE_CLASS: 'grey69',
+}
+
+LOG_THEME = {
+    f"{ReprHighlighter.base_style}{doc_type}": f"{doc_style} bold"
+    for doc_type, doc_style in DOC_TYPE_STYLES.items()
+}
+
+LOG_THEME[f"{ReprHighlighter.base_style}epstein_filename"] = FILENAME_STYLE
+LOG_LEVEL_ENV_VAR = 'LOG_LEVEL'
+
+
+class LogHighlighter(ReprHighlighter):
+    highlights = ReprHighlighter.highlights + [
+        *[fr"(?P<{doc_type}>{doc_type})" for doc_type in DOC_TYPE_STYLES.keys()],
+        "(?P<epstein_filename>" + FILE_NAME_REGEX.pattern + ')',
+    ]
+
+
+log_console = Console(color_system='256', theme=Theme(LOG_THEME))
+log_handler = RichHandler(console=log_console, highlighter=LogHighlighter())
+logging.basicConfig(level="NOTSET", format="%(message)s", datefmt="[%X]", handlers=[log_handler])
+logger = logging.getLogger("rich")
+
+
+# Log levels
+datefinder_logger = logging.getLogger('datefinder')  # Set log level to suppress annoying output
+env_log_level_str = environ.get(LOG_LEVEL_ENV_VAR) or None
+env_log_level = None
+
+if env_log_level_str:
+    try:
+        env_log_level = getattr(logging, env_log_level_str)
+    except Exception as e:
+        logger.warning(f"{LOG_LEVEL_ENV_VAR}='{env_log_level_str}' does not exist, defaulting to DEBUG")
+        env_log_level = logging.DEBUG
+
+    logger.warning(f"Setting log level to {env_log_level} based on {LOG_LEVEL_ENV_VAR} env var...")
+    logger.setLevel(env_log_level)
+    datefinder_logger.setLevel(env_log_level)

epstein_files/util/rich.py

@@ -18,9 +18,11 @@ from epstein_files.util.constant.names import UNKNOWN
 from epstein_files.util.constant.strings import DEFAULT, EMAIL, NA, OTHER_SITE_LINK_STYLE, QUESTION_MARKS, SiteType
 from epstein_files.util.constant.urls import *
 from epstein_files.util.constants import FALLBACK_TIMESTAMP, HEADER_ABBREVIATIONS
-from epstein_files.util.env import args, logger
+from epstein_files.util.data import json_safe
+from epstein_files.util.env import args
 from epstein_files.util.file_helper import file_size_str
-from epstein_files.util.highlighted_group import HIGHLIGHTED_GROUPS, InterestingNamesHighlighter
+from epstein_files.util.highlighted_group import ALL_HIGHLIGHTS, HIGHLIGHTED_NAMES, EpsteinHighlighter
+from epstein_files.util.logging import logger
 
 TITLE_WIDTH = 50
 NUM_COLOR_KEY_COLS = 4
@@ -38,20 +40,19 @@ TITLE_STYLE = 'black on bright_white bold'
 
 HIGHLIGHTED_GROUP_COLOR_KEYS = [
     Text(highlight_group.label.replace('_', ' '), style=highlight_group.style)
-    for highlight_group in sorted(HIGHLIGHTED_GROUPS, key=lambda hg: hg.label)
-    if not highlight_group.is_multiline
+    for highlight_group in sorted(HIGHLIGHTED_NAMES, key=lambda hg: hg.label)
 ]
 
 THEME_STYLES = {
     DEFAULT: 'wheat4',
     TEXT_LINK: 'deep_sky_blue4 underline',
-    **{hg.theme_style_name: hg.style for hg in HIGHLIGHTED_GROUPS},  # Inject style names for HighlightedGroups
+    **{hg.theme_style_name: hg.style for hg in ALL_HIGHLIGHTS},
 }
 
 # Instantiate console object
 CONSOLE_ARGS = {
     'color_system': '256',
-    'highlighter': InterestingNamesHighlighter(),
+    'highlighter': EpsteinHighlighter(),
     'record': args.build,
     'safe_box': False,
     'theme': Theme(THEME_STYLES),
@@ -121,7 +122,7 @@ def print_centered_link(url: str, link_text: str, style: str | None = None) -> N
     print_centered(link_markup(url, link_text, style or ARCHIVE_LINK_COLOR))
 
 
-def print_color_key(_key_type: Literal["Groups", "People"] = "Groups") -> None:
+def print_color_key() -> None:
     color_table = Table(title=f'Rough Guide to Highlighted Colors', show_header=False)
     num_colors = len(HIGHLIGHTED_GROUP_COLOR_KEYS)
     row_number = 0
@@ -153,13 +154,13 @@ def print_header(epstein_files: 'EpsteinFiles') -> None:
 
 
 def print_json(label: str, obj: object, skip_falsey: bool = False) -> None:
+    print(obj)
+
     if isinstance(obj, dict):
         if skip_falsey:
             obj = {k: v for k, v in obj.items() if v}
 
-        if None in obj:
-            obj = {k or UNKNOWN: v for k, v in obj.items()}
-
+        obj = json_safe(obj)
 
     console.line()
     console.print(Panel(label, expand=False))
@@ -219,6 +220,10 @@ def print_other_site_link(is_header: bool = True) -> None:
     word_count_link = link_text_obj(WORD_COUNT_URL, 'site showing the most frequently used words in these communiques', OTHER_SITE_LINK_STYLE)
     print_centered(parenthesize(word_count_link))
 
+    if is_header:
+        metadata_link = link_text_obj(JSON_METADATA_URL, 'file metadata containing author attribution explanations', OTHER_SITE_LINK_STYLE)
+        print_centered(parenthesize(metadata_link))
+
 
 def print_page_title(expand: bool = True, width: int | None = None) -> None:
     title_panel = Panel(Text(PAGE_TITLE, justify='center'), expand=expand, style=TITLE_STYLE, width=width)
@@ -292,7 +297,7 @@ def write_html(output_path: Path) -> None:
         logger.warning(f"Not writing HTML because args.build={args.build}.")
         return
 
-    console.save_html(output_path, code_format=CONSOLE_HTML_FORMAT, theme=HTML_TERMINAL_THEME)
+    console.save_html(str(output_path), code_format=CONSOLE_HTML_FORMAT, theme=HTML_TERMINAL_THEME)
     logger.warning(f"Wrote {file_size_str(output_path)} to '{output_path}'")
 
 
@@ -320,5 +325,5 @@ def _print_external_links() -> None:
     print_centered(link_markup(EPSTEIN_WEB_URL) + " (character summaries)")
 
 
-if args.deep_debug:
-    print_json('THEME_STYLES', THEME_STYLES)
+# if args.deep_debug:
+#     print_json('THEME_STYLES', THEME_STYLES)

epstein_files/util/search_result.py

@@ -1,15 +1,23 @@
 from dataclasses import dataclass
 
 from rich.text import Text
+# from epstein_files.documents.document import type Document
 
-from epstein_files.documents.document import Document
+
+@dataclass
+class MatchedLine:
+    line: str
+    line_number: int
+
+    def __rich__(self) -> Text:
+        return Text('').append(str(self.line_number), style='cyan').append(f":{self.line}")
+
+    def __str__(self) -> str:
+        return f"{self.line_number}:{self.line}"
 
 
 @dataclass
 class SearchResult:
     """Simple class used for collecting documents that match a given search term."""
-    document: Document
-    lines: list[Text]  # The lines that match the search
-
-    def unprefixed_lines(self) -> list[str]:
-        return [line.plain.split(':', 1)[1] for line in self.lines]
+    document: 'Document'
+    lines: list[MatchedLine]

epstein_files/util/timer.py

@@ -0,0 +1,24 @@
+import time
+from dataclasses import dataclass, field
+
+from epstein_files.util.logging import logger
+
+
+@dataclass
+class Timer:
+    started_at: float = field(default_factory=lambda: time.perf_counter())
+    checkpoint_at: float = field(default_factory=lambda: time.perf_counter())
+    decimals: int = 2
+
+    def print_at_checkpoint(self, msg: str) -> None:
+        logger.warning(f"{msg} in {self.seconds_since_checkpoint_str()}")
+        self.checkpoint_at = time.perf_counter()
+
+    def seconds_since_checkpoint_str(self) -> str:
+        return f"{(time.perf_counter() - self.checkpoint_at):.{self.decimals}f} seconds"
+
+    def seconds_since_start(self) -> float:
+        return time.perf_counter() - self.started_at
+
+    def seconds_since_start_str(self) -> str:
+        return f"{self.seconds_since_start():.{self.decimals}f} seconds"

epstein_files/util/word_count.py

@@ -12,7 +12,8 @@ from epstein_files.documents.emails.email_header import EmailHeader
 from epstein_files.util.constant.common_words import COMMON_WORDS, UNSINGULARIZABLE_WORDS
 from epstein_files.util.constant.names import OTHER_NAMES
 from epstein_files.util.data import ALL_NAMES, flatten, sort_dict
-from epstein_files.util.env import args, logger
+from epstein_files.util.env import args
+from epstein_files.util.logging import logger
 from epstein_files.util.rich import highlighter
 from epstein_files.util.search_result import SearchResult
 

{epstein_files-1.0.0.dist-info → epstein_files-1.0.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: epstein-files
-Version: 1.0.0
+Version: 1.0.1
 Summary: Tools for working with the Jeffrey Epstein documents released in November 2025.
 Author: Michel de Cryptadamus
 Requires-Python: >=3.11,<4.0
@@ -20,7 +20,8 @@ Description-Content-Type: text/markdown
 * The Epstein text messages (and some of the emails along with summary counts of sent emails to/from Epstein) generated by this code can be viewed [here](https://michelcrypt4d4mus.github.io/epstein_text_messages/).
 * All of His Emails can be read at another page also generated by this code [here](https://michelcrypt4d4mus.github.io/epstein_emails_house_oversight/).
 * Word counts for the emails and text messages are [here](https://michelcrypt4d4mus.github.io/epstein_text_messages/epstein_emails_word_count.html).
-* Configuration variables assigning specific `HOUSE_OVERSIGHT_XXXXXX.txt` file IDs (the `111111` part) as being emails to or from particular people based on various research and contributions can be found in the [constants.py](./epstein_files/util/constants.py) file in this repo.
+* Metadata containing what I have figured out about who sent or received the communications in a given file (and a brief explanation for how I figured it out for each file) is deployed [here](https://michelcrypt4d4mus.github.io/epstein_text_messages/epstein_files_nov_2025_cryptadamus_metadata.json)
+* Configuration variables assigning specific `HOUSE_OVERSIGHT_XXXXXX.txt` file IDs (the `111111` part) as being emails to or from particular people based on various research and contributions can be found in [constants.py](./epstein_files/util/constants.py). Everything in `constants.py` should also appear in the JSON metadata.
 
 
 ### Usage

epstein_files-1.0.1.dist-info/RECORD

@@ -0,0 +1,30 @@
+epstein_files/__init__.py,sha256=HKOGWVzyiEl0II8s7qnO33wcTZPw8dr3HDfT7_VnY2U,7831
+epstein_files/documents/communication.py,sha256=SunZdjMhR9v6y8LlQ6jhIu8vYjSndaBK0Su1mKnhfj0,2060
+epstein_files/documents/document.py,sha256=E2VUuHSoam8r8pyRbWy68_cVFU9DJ0irbtYULaDBYzE,16378
+epstein_files/documents/email.py,sha256=IpKeOuLTmHWowBvUUEp-tyTC8pwEmXg3XgLJkplQZWg,37717
+epstein_files/documents/emails/email_header.py,sha256=wkPfSLbmzkAeQwvhf0bAeFDLPbQT-EeG0v8vNNLYktM,7502
+epstein_files/documents/imessage/text_message.py,sha256=DMdd__L7UPad0YS49MJf_3bTVyE1BLfWafQbDQierC8,3272
+epstein_files/documents/json_file.py,sha256=Vf9iHiEhUqk-hEh6lGZX_mJNWZUUQWeT-UI3FhKW0hw,1101
+epstein_files/documents/messenger_log.py,sha256=-hSMFH_dyrjGLdph4SU2xQK4RpXOhkR3R_nqRrz3e1w,5620
+epstein_files/documents/other_file.py,sha256=JbKDtNrQ1Ua5vGPlZayON5Kgy0oJ-gHcdO9l9Iv9wRo,8930
+epstein_files/epstein_files.py,sha256=ruhl24FgOpEdcqbMcIpgyYlqEBi1ZeztQ7QmZMYANp0,17856
+epstein_files/util/constant/common_words.py,sha256=aR0UjoWmxyR49XS-DtHECQ1CiA_bK8hNP6CQ1TS9yZA,3696
+epstein_files/util/constant/html.py,sha256=9U098TGzlghGg4WfxLYHyub5JGR17Dv7VP5i2MSu8Kk,1415
+epstein_files/util/constant/names.py,sha256=eSgHnOqMAy7Czfw1vJ3lvniDST7QpfuQ8XgzttXpAi8,10191
+epstein_files/util/constant/strings.py,sha256=3JTqD0luJrC3NbGXn4q6P-gIaaNVx36P1oCmp92gAoM,1750
+epstein_files/util/constant/urls.py,sha256=P_nnb5cIrJ2NCbd4CoRfilZ5jrOzFs30jNK50F7IjZw,5483
+epstein_files/util/constants.py,sha256=l0TrpncCJcecar5geggLVfOJVhicWcu3zws-62Cslzg,102472
+epstein_files/util/data.py,sha256=s7EgmQ6SC6MoUMhP8GaxglJOJpU_nRmwI-T2inLl8hs,3329
+epstein_files/util/doc_cfg.py,sha256=AFLBp9PO2tBUYhTeNztpkWs69VPr41fCuA22IesM59g,9914
+epstein_files/util/env.py,sha256=9wM4mKPZoJO9J5VKLP67ZR1IEMkH2UaGdIaYm2tptMQ,4581
+epstein_files/util/file_helper.py,sha256=ea428og2gyUsMWd2GFldmAGvuqGNNHaTc7gGGSM_0aI,3827
+epstein_files/util/highlighted_group.py,sha256=UO0pqQH29JHiOMK2psIozIeA0zinDkql-NYDGRmFUzA,35173
+epstein_files/util/logging.py,sha256=0mABrI5TvodLDe-9svHv3YaIdSyr8PXFcu4nntW1WsI,1858
+epstein_files/util/rich.py,sha256=cLkiSg3f68bUxobzRiBRfrgS7Ya-FK7ltpQcZXNHMX4,13317
+epstein_files/util/search_result.py,sha256=1fxe0KPBQXBk4dLfu6m0QXIzYfZCzvaSkWqvghJGzxY,567
+epstein_files/util/timer.py,sha256=8hxW4Y1JcTUfnBrHh7sL2pM9xu1sL4HFQM4CmmzTarU,837
+epstein_files/util/word_count.py,sha256=XTINgLm01jFQlNgdiLCcVFCodXAIb1dNbaAvznoRb1o,6757
+epstein_files-1.0.1.dist-info/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
+epstein_files-1.0.1.dist-info/METADATA,sha256=IK5B-ltx_L0-EK2Oq2Q2gVXWlhYqKoYw6HCFisqitvQ,3587
+epstein_files-1.0.1.dist-info/WHEEL,sha256=d2fvjOD7sXsVzChCqf0Ty0JbHKBaLYwDbGQDwQTnJ50,88
+epstein_files-1.0.1.dist-info/RECORD,,

epstein_files-1.0.0.dist-info/RECORD

@@ -1,28 +0,0 @@
-epstein_files/__init__.py,sha256=JlGEliI8H5TqdGuCTtGIwLiM7GutbyND-IURG5Z-gAA,7442
-epstein_files/documents/communication.py,sha256=XybnJDqyK00RTeEkbLWgrVDkZS7pmC8EoftX1_G4OIM,2066
-epstein_files/documents/document.py,sha256=6av_KLTm70KIABl9Fl7DL9jKkvqdN18OLmECj2PAkEs,15223
-epstein_files/documents/email.py,sha256=rzgyVMBlKo9ytRdE6Eif_8zF4fecaMduIgPQkvRmmk0,34283
-epstein_files/documents/emails/email_header.py,sha256=JqcSqa-amdombgszE_cjXBrvZKE0HxU2mNAoD44Ekh4,7479
-epstein_files/documents/imessage/text_message.py,sha256=zU46FHeHKG6vbIZ-qLDdZC-1f3VmXW73fKo-BKOCgq8,3199
-epstein_files/documents/json_file.py,sha256=sL99E_34X9BIwQ27vE0Afy446oiyCmPLYdsJ6-OGMHU,611
-epstein_files/documents/messenger_log.py,sha256=HbtqthtkzXhkY7DxwxFOBy6xx6gPmmXIxZ4bsloLHEg,2995
-epstein_files/documents/other_file.py,sha256=_Z8EVLLy1u8RuEr5iPTKJdYZjQHFHp0o-mBErBYsRXg,4699
-epstein_files/epstein_files.py,sha256=Fb-SPwRlHZPB7iL48dL3Da4nPyzF9_Rk5nNpVnfku2U,20131
-epstein_files/util/constant/common_words.py,sha256=aR0UjoWmxyR49XS-DtHECQ1CiA_bK8hNP6CQ1TS9yZA,3696
-epstein_files/util/constant/html.py,sha256=9U098TGzlghGg4WfxLYHyub5JGR17Dv7VP5i2MSu8Kk,1415
-epstein_files/util/constant/names.py,sha256=_iNHJsr4sdUXG1cozllRy8yV6SYXgsr6oTPHTCbrpj0,9967
-epstein_files/util/constant/strings.py,sha256=B5RpaGCc1EqiDKF-3BabNy4dJU6n4kXytx96aanHIDA,1115
-epstein_files/util/constant/urls.py,sha256=pks5frfWXmFwEkfB33Uy7IR1mn0DaAI_8kDjA_g03yg,5089
-epstein_files/util/constants.py,sha256=jErrj2r_K8pbFNih7WxlcP0sPO7ZyzrLxH3MlSw5xR4,106137
-epstein_files/util/data.py,sha256=7yz3osIlpjV9nbPEvRBmExokBsgm20kitSfqNpaYWj4,3746
-epstein_files/util/env.py,sha256=T-sd1lPfw-L6WXvpIv-cqGTqBfZvacZvniW1OKOdkrU,4624
-epstein_files/util/file_cfg.py,sha256=E7VVxVZ_cSN18HgI6q8e6UwTApMWuPC9qTqb3_tXCqs,7183
-epstein_files/util/file_helper.py,sha256=EHVRFxcWmc0RD1VUeLFCcmM-82c5FISIUBIwg4TodsY,2727
-epstein_files/util/highlighted_group.py,sha256=0MpzBYVaGfyUZQwFP7EVrLHDYU57qnnmNDC56gVxq2Q,33091
-epstein_files/util/rich.py,sha256=Z28-etKpZdc6aPxdmAxNe9ZvMWK28iubYASt50VxtJ4,13194
-epstein_files/util/search_result.py,sha256=5nkR-YzUiR52FmGIad_Uw0XaU2E_FfHZBHgc8dm3sAs,429
-epstein_files/util/word_count.py,sha256=L1T6zjShaPZo1ilpM0dSAT3X-EgaZ4YEigHNLxjKit4,6719
-epstein_files-1.0.0.dist-info/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
-epstein_files-1.0.0.dist-info/METADATA,sha256=dGc412CXHc2EHOd0C7HuEa2qFQFh42dNbDw5CiH0Ny4,3239
-epstein_files-1.0.0.dist-info/WHEEL,sha256=d2fvjOD7sXsVzChCqf0Ty0JbHKBaLYwDbGQDwQTnJ50,88
-epstein_files-1.0.0.dist-info/RECORD,,