epstein-files 1.0.12__py3-none-any.whl → 1.0.14__py3-none-any.whl

epstein_files/util/highlighted_group.py

@@ -299,23 +299,10 @@ HIGHLIGHTED_NAMES = [
             KEN_STARR: 'head of the Monica Lewinsky investigation against Bill Clinton',
         }
     ),
-    HighlightedNames(
-        label='friend',
-        style='tan',
-        pattern=r"Andrew Farkas|Thomas\s*(J\.?\s*)?Barrack(\s*Jr)?",
-        emailers = {
-            DAVID_STERN: f'emailed Epstein from Moscow, appears to know chairman of {DEUTSCHE_BANK}',
-            JONATHAN_FARKAS: "heir to the Alexander's department store fortune",
-            'linkspirit': "Skype username of someone Epstein communicated with",
-            'Peter Thomas Roth': 'student of Epstein at Dalton, skincare company founder',
-            STEPHEN_HANSON: None,
-            TOM_BARRACK: 'long time friend of Trump',
-        }
-    ),
     HighlightedNames(
         label=FINANCE,
         style='green',
-        pattern=r'Apollo|Ari\s*Glass|Bank|(Bernie\s*)?Madoff|Black(rock|stone)|B\s*of\s*A|Boothbay(\sFund\sManagement)?|Chase\s*Bank|Credit\s*Suisse|DB|Deutsche\s*(Asset|Bank)|Electron\s*Capital\s*(Partners)?|Fenner|FRBNY|Goldman(\s*Sachs)|HSBC|Invesco|(Janet\s*)?Yellen|(Jerome\s*)?Powell(?!M\. Cabot)|(Jimmy\s*)?Cayne|JPMC?|j\.?p\.?\s*morgan(\.?com|\s*Chase)?|Madoff|Merrill(\s*Lynch)?|(Michael\s*)?(Cembalest|Milken)|Mizrahi\s*Bank|MLPF&S|((anti.?)?money\s+)?launder(s?|ers?|ing)?(\s+money)?|Morgan Stanley|(Peter L. )?Scher|(Ray\s*)?Dalio|Schwartz?man|Serageldin|UBS|us.gio@jpmorgan.com',
+        pattern=r'Apollo|Ari\s*Glass|Bank|(Bernie\s*)?Madoff|Black(rock|stone)|B\s*of\s*A|Boothbay(\sFund\sManagement)?|Chase\s*Bank|Credit\s*Suisse|DB|Deutsche\s*(Asset|Bank)|Electron\s*Capital\s*(Partners)?|Fenner|FRBNY|Goldman(\s*Sachs)|HSBC|Invesco|(Janet\s*)?Yellen|(Jerome\s*)?Powell(?!M\. Cabot)|(Jimmy\s*)?Cayne|JPMC?|j\.?p\.?\s*morgan(\.?com|\s*Chase)?|Madoff|Merrill(\s*Lynch)?|(Michael\s*)?(Cembalest|Milken)|Mizrahi\s*Bank|MLPF&S|((anti.?)?money\s+)?launder(s?|ers?|ing)?(\s+money)?|Morgan Stanley|(Peter L. )?Scher|(Ray\s*)?Dalio|(Richard\s*)?LeFrak|Schwartz?man|Serageldin|UBS|us.gio@jpmorgan.com',
         emailers={
             AMANDA_ENS: 'Citigroup',
             BRAD_WECHSLER: f"head of {LEON_BLACK}'s personal investment vehicle according to FT",
@@ -332,9 +319,23 @@ HIGHLIGHTED_NAMES = [
             PAUL_MORRIS: 'Deutsche Bank',
         }
     ),
+    HighlightedNames(
+        label='friend',
+        style='tan',
+        pattern=r"Andrew Farkas|Thomas\s*(J\.?\s*)?Barrack(\s*Jr)?",
+        emailers = {
+            DANGENE_AND_JENNIE_ENTERPRISE: 'founders of the members-only CORE club',
+            DAVID_STERN: f'emailed Epstein from Moscow, appears to know chairman of {DEUTSCHE_BANK}',
+            JONATHAN_FARKAS: "heir to the Alexander's department store fortune",
+            'linkspirit': "Skype username of someone Epstein communicated with",
+            'Peter Thomas Roth': 'student of Epstein at Dalton, skincare company founder',
+            STEPHEN_HANSON: None,
+            TOM_BARRACK: 'long time friend of Trump',
+        },
+    ),
     HighlightedNames(
         label=HARVARD.lower(),
-        style='deep_pink2',
+        style='light_goldenrod3',
         pattern=r'Cambridge|(Derek\s*)?Bok|Elisa(\s*New)?|Harvard(\s*(Business|Law|University)(\s*School)?)?|(Jonathan\s*)?Zittrain|(Stephen\s*)?Kosslyn',
         emailers = {
             "Donald Rubin": f"Professor of Statistics",
@@ -372,14 +373,6 @@ HIGHLIGHTED_NAMES = [
         style='color(168)',
         pattern=r'BOJ|(Bank\s+of\s+)?Japan(ese)?|jpy?(?! Morgan)|SG|Singapore|Toky[op]',
     ),
-    HighlightedNames(
-        label='javanka',
-        style='medium_violet_red',
-        emailers = {
-            IVANKA: None,
-            JARED_KUSHNER: None,
-        }
-    ),
     HighlightedNames(
         label=JOURNALIST,
         style='bright_yellow',
@@ -403,7 +396,7 @@ HIGHLIGHTED_NAMES = [
     HighlightedNames(
         label='law enforcement',
         style='color(24) bold',
-        pattern=r'ag|(Alicia\s*)?Valle|AML|attorney|((Bob|Robert)\s*)?Mueller|(Byung\s)?Pak|CFTC?|CIA|CIS|CVRA|Dep(artmen)?t\.?\s*of\s*(the\s*)?(Justice|Treasury)|DHS|DOJ|FBI|FCPA|FDIC|Federal\s*Bureau\s*of\s*Investigation|FinCEN|FINRA|FOIA|FTC|IRS|(James\s*)?Comey|(Jennifer\s*Shasky\s*)?Calvery|((Judge|Mark)\s*)?(Carney|Filip)|(Kirk )?Blouin|KYC|NIH|NS(A|C)|OCC|OFAC|(Lann?a\s*)?Belohlavek|lawyer|(Michael\s*)?Reiter|OGE|Office\s*of\s*Government\s*Ethics|Police Code Enforcement|(Preet\s*)?Bharara|SCOTUS|SD(FL|NY)|Southern\s*District\s*of\s*(Florida|New\s*York)|SEC|Secret\s*Service|Securities\s*and\s*Exchange\s*Commission|State\s*Dep(artmen)?t|Strzok|Supreme\s*Court|Treasury\s*(Dep(artmen)?t|Secretary)|TSA|USAID|(William\s*J\.?\s*)?Zloch',
+        pattern=r'ag|(Alicia\s*)?Valle|AML|(Andrew\s*)?McCabe|attorney|((Bob|Robert)\s*)?Mueller|(Byung\s)?Pak|CFTC?|CIA|CIS|CVRA|Dep(artmen)?t\.?\s*of\s*(the\s*)?(Justice|Treasury)|DHS|DOJ|FBI|FCPA|FDIC|Federal\s*Bureau\s*of\s*Investigation|FinCEN|FINRA|FOIA|FTC|IRS|(James\s*)?Comey|(Jennifer\s*Shasky\s*)?Calvery|((Judge|Mark)\s*)?(Carney|Filip)|(Kirk )?Blouin|KYC|NIH|NS(A|C)|OCC|OFAC|(Lann?a\s*)?Belohlavek|lawyer|(Michael\s*)?Reiter|OGE|Office\s*of\s*Government\s*Ethics|Police Code Enforcement|(Preet\s*)?Bharara|SCOTUS|SD(FL|NY)|SEC|Secret\s*Service|Securities\s*and\s*Exchange\s*Commission|Southern\s*District\s*of\s*(Florida|New\s*York)|State\s*Dep(artmen)?t|Strzok|Supreme\s*Court|Treasury\s*(Dep(artmen)?t|Secretary)|TSA|USAID|(William\s*J\.?\s*)?Zloch',
         emailers = {
             ANN_MARIE_VILLAFANA: 'southern district of Florida U.S. Attorney',
             DANNY_FROST: 'Director of Communications at Manhattan DA',
@@ -462,6 +455,7 @@ HIGHLIGHTED_NAMES = [
             CHRISTINA_GALBRAITH: f"{REPUTATION_MGMT}, worked on Epstein's Google search results with {TYLER_SHEARS}",
             IAN_OSBORNE: f"{OSBORNE_LLP} reputation repairer possibly hired by Epstein ca. 2011-06",
             MICHAEL_SITRICK: 'crisis PR',
+            'Owen Blicksilver': 'OBPR, Inc.',
             PEGGY_SIEGAL: 'socialite',
             'R. Couri Hay': None,
             ROSS_GOW: 'Acuity Reputation Management',
@@ -471,10 +465,10 @@ HIGHLIGHTED_NAMES = [
     HighlightedNames(
         label='republicans',
         style='bold dark_red',
-        pattern=r'Alberto\sGonzale[sz]|(Alex\s*)?Acosta|(Bill\s*)?Barr|Bill\s*Shine|(Bob\s*)?Corker|(John\s*(R.?\s*)?)Bolton|Broidy|(Chris\s)?Christie|Devin\s*Nunes|(Don\s*)?McGa[hn]n|McMaster|(George\s*)?Nader|GOP|(Brett\s*)?Kavanaugh|Kissinger|Kobach|Koch\s*Brothers|Kolfage|Kudlow|Lewandowski|(Marco\s)?Rubio|(Mark\s*)Meadows|Mattis|McCain|(?<!Merwin Dela )Cruz|(Michael\s)?Hayden|((General|Mike)\s*)?(Flynn|Pence)|(Mitt\s*)?Romney|Mnuchin|Nikki|Haley|(Paul\s+)?(Manafort|Volcker)|(Peter\s)?Navarro|Pompeo|Reagan|Reince|Priebus|Republican|(Rex\s*)?Tillerson|(?<!Cynthia )(Richard\s*)?Nixon|Sasse',
+        pattern=r'Alberto\sGonzale[sz]|(Alex\s*)?Acosta|(Bill\s*)?Barr|Bill\s*Shine|(Bob\s*)?Corker|(John\s*(R.?\s*)?)Bolton|Broidy|(Chris\s)?Christie|Devin\s*Nunes|(Don\s*)?McGa[hn]n|McMaster|(George\s*)?Nader|GOP|(Brett\s*)?Kavanaugh|Kissinger|Kobach|Koch\s*Brothers|Kolfage|Kudlow|Lewandowski|(Marco\s)?Rubio|(Mark\s*)Meadows|Mattis|McCain|(?<!Merwin Dela )Cruz|(Michael\s)?Hayden|((General|Mike)\s*)?(Flynn|Pence)|(Mitt\s*)?Romney|Mnuchin|Nikki|Haley|(Paul\s+)?(Manafort|Volcker)|(Peter\s)?Navarro|Pompeo|Reagan|Reince|Priebus|Republican|(Rex\s*)?Tillerson|(?<!Cynthia )(Richard\s*)?Nixon|Sasse|Tea\s*Party',
         # There's no emails from these people, they're just here to automate the regex creation for both first + last names
         emailers = {
-            RUDY_GIULIANI: 'disbarred formed mayor of New York City',
+            RUDY_GIULIANI: None,
             TULSI_GABBARD: None,
         },
     ),
@@ -482,7 +476,7 @@ HIGHLIGHTED_NAMES = [
         label='Rothschild family',
         style='indian_red',
         emailers={
-            ARIANE_DE_ROTHSCHILD: None,
+            ARIANE_DE_ROTHSCHILD: 'heiress',
             JOHNNY_EL_HACHEM: f'Works with {ARIANE_DE_ROTHSCHILD}',
         },
     ),
@@ -502,7 +496,7 @@ HIGHLIGHTED_NAMES = [
         style='light_goldenrod2',
         pattern=r'Alain Forget|Brotherton|Carl\s*Sagan|Columbia|David Grosof|J(ames|im)\s*Watson|(Lord\s*)?Martin\s*Rees|Massachusetts\s*Institute\s*of\s*Technology|MIT(\s*Media\s*Lab)?|Media\s*Lab|Minsky|((Noam|Valeria)\s*)?Chomsky|Norman\s*Finkelstein|Praluent|Regeneron|(Richard\s*)?Dawkins|Sanofi|Stanford|(Stephen\s*)?Hawking|(Steven?\s*)?Pinker|UCLA',
         emailers = {
-            DAVID_HAIG: None,
+            DAVID_HAIG: 'evolutionary geneticist?',
             JOSCHA_BACH: 'cognitive science / AI research',
             'Daniel Kahneman': 'Nobel economic sciences laureate and cognitivie psychologist (?)',
             'Ed Boyden': 'Associate Professor, MIT Media Lab neurobiology',
@@ -536,7 +530,7 @@ HIGHLIGHTED_NAMES = [
     HighlightedNames(
         label='trump',
         style='red3 bold',
-        pattern=r"@?realDonaldTrump|(Alan\s*)?Weiss?elberg|\bDJ?T\b|Donald J. Tramp|(Donald\s+(J\.\s+)?)?Trump(ism|\s*Properties)?|Don(ald| *Jr)(?! Rubin)|Ivana|(Madeleine\s*)?Westerhout|Mar[-\s]*a[-\s]*Lago|(Marla\s*)?Maples|(Matt(hew)? )?Calamari|\bMatt C\b|Melania|(Michael (J.? )?)?Boccio|Rebekah\s*Mercer|Roger\s+Stone|rona|(The\s*)?Art\s*of\s*the\s*Deal",
+        pattern=r"@?realDonaldTrump|(Alan\s*)?Weiss?elberg|\bDJ?T\b|Donald J. Tramp|(Donald\s+(J\.\s+)?)?Trump(ism|\s*Properties)?|Don(ald| *Jr)(?! Rubin)|Ivank?a|Jared|Kushner|(Madeleine\s*)?Westerhout|Mar[-\s]*a[-\s]*Lago|(Marla\s*)?Maples|(Matt(hew)? )?Calamari|\bMatt C\b|Melania|(Michael (J.? )?)?Boccio|Rebekah\s*Mercer|Roger\s+Stone|rona|(The\s*)?Art\s*of\s*the\s*Deal",
         emailers = {
             'Bruce Moskowitz': "'Trump's health guy' according to Epstein",
         },
@@ -558,7 +552,7 @@ HIGHLIGHTED_NAMES = [
     HighlightedNames(
         label=VIRGIN_ISLANDS,
         style='sea_green1',
-        pattern=r'Antigua|Bahamas|Caribb?ean|Dominican\s*Republic|(Great|Little)\s*St.?\s*James|Haiti(an)?|(John\s*)deJongh(\s*Jr\.?)|(Kenneth E\. )?Mapp|Palm\s*Beach(?!\s*Post)|PBI|S(ain)?t.?\s*Thomas|USVI|VI|(The\s*)?Virgin\s*Islands(\s*Daily\s*News)?',  # TODO: VI Daily News should be yellow but it's hard bc Daily News xists
+        pattern=r'Antigua|Bahamas|Caribb?ean|Dominican\s*Republic|(Great|Little)\s*St.?\s*James|Haiti(an)?|(John\s*)deJongh(\s*Jr\.?)|(Kenneth E\. )?Mapp|Palm\s*Beach(?!\s*Post)|PBI|S(ain)?t.?\s*Thomas|USVI|(?<!Epstein )VI|(The\s*)?Virgin\s*Islands(\s*Daily\s*News)?',  # TODO: VI Daily News should be yellow but it's hard bc Daily News xists
         emailers = {
             CECILE_DE_JONGH: f'First lady 2007-2015',
             STACEY_PLASKETT: 'non-voting member of Congress',
@@ -667,7 +661,7 @@ def get_info_for_name(name: str) -> str | None:
 def get_style_for_category(category: str) -> str | None:
     if category in CATEGORY_STYLES:
         return CATEGORY_STYLES[category]
-    elif category in [CONFERENCE, SPEECH]:
+    elif category == CONFERENCE:
         return f"{get_style_for_category(ACADEMIA)} dim"
     elif category == SOCIAL:
         return get_style_for_category(PUBLICIST)

epstein_files/util/output.py

@@ -32,8 +32,6 @@ DEFAULT_EMAILERS = [
     MASHA_DROKOVA,
     RENATA_BOLOTOVA,
     STEVE_BANNON,
-    OLIVIER_COLOM,
-    BORIS_NIKOLIC,
     PRINCE_ANDREW,
     JIDE_ZEITLIN,
     DAVID_STERN,
@@ -60,7 +58,7 @@ if len(set(DEFAULT_EMAILERS).intersection(set(DEFAULT_EMAILER_TABLES))) > 0:
 def print_emails(epstein_files: EpsteinFiles) -> int:
     """Returns number of emails printed."""
     print_section_header(('Selections from ' if not args.all_emails else '') + 'His Emails')
-    print_other_site_link(is_header=False)
+    print_all_files_page_link(epstein_files)
     emailers_to_print: list[str | None]
     emailer_tables: list[str | None] = []
     already_printed_emails: list[Email] = []
@@ -69,7 +67,7 @@ def print_emails(epstein_files: EpsteinFiles) -> int:
     if specified_names:
         emailers_to_print = specified_names
     else:
-        epstein_files.print_emailer_counts_table()
+        print_centered(Padding(epstein_files.table_of_emailers(), (2, 0)))
 
         if args.all_emails:
             emailers_to_print = sorted(epstein_files.all_emailers(), key=lambda e: epstein_files.earliest_email_at(e))
@@ -105,8 +103,8 @@ def print_emails(epstein_files: EpsteinFiles) -> int:
     if args.all_emails:
         _verify_all_emails_were_printed(epstein_files, already_printed_emails)
 
-    fwded_articles = [e for e in already_printed_emails if e.config and e.config.is_fwded_article]
-    log_msg = f"Rewrote {len(Email.rewritten_header_ids)} email headers (out of {len(already_printed_emails)})"
+    fwded_articles = [e for e in already_printed_emails if e.config and e.is_fwded_article()]
+    log_msg = f"Rewrote {len(Email.rewritten_header_ids)} of {len(already_printed_emails)} email headers"
     logger.warning(f"{log_msg}, {len(fwded_articles)} of the emails were forwarded articles.")
     return len(already_printed_emails)
 
@@ -125,17 +123,6 @@ def print_json_files(epstein_files: EpsteinFiles):
             console.print_json(json_file.json_str(), indent=4, sort_keys=False)
 
 
-def write_json_metadata(epstein_files: EpsteinFiles) -> None:
-    json_str = epstein_files.json_metadata()
-
-    if args.build:
-        with open(JSON_METADATA_PATH, 'w') as f:
-            f.write(json_str)
-            log_file_write(JSON_METADATA_PATH)
-    else:
-        console.print_json(json_str, indent=4, sort_keys=True)
-
-
 def print_json_stats(epstein_files: EpsteinFiles) -> None:
     console.line(5)
     console.print(Panel('JSON Stats Dump', expand=True, style='reverse bold'), '\n')
@@ -149,21 +136,19 @@ def print_json_stats(epstein_files: EpsteinFiles) -> None:
     print_json("count_by_month", count_by_month(epstein_files.all_documents()))
 
 
-def print_text_messages(epstein_files: EpsteinFiles) -> None:
-    print_section_header('Text Messages')
-    print_centered("(conversations are sorted chronologically based on timestamp of first message)\n", style='gray30')
-    authors: list[str | None] = specified_names if specified_names else [JEFFREY_EPSTEIN]
-    log_files = epstein_files.imessage_logs_for(authors)
-
-    for log_file in log_files:
-        console.print(Padding(log_file))
-        console.line(2)
+def write_json_metadata(epstein_files: EpsteinFiles) -> None:
+    json_str = epstein_files.json_metadata()
 
-    epstein_files.print_imessage_summary()
+    if args.build:
+        with open(JSON_METADATA_PATH, 'w') as f:
+            f.write(json_str)
+            log_file_write(JSON_METADATA_PATH)
+    else:
+        console.print_json(json_str, indent=4, sort_keys=True)
 
 
 def write_urls() -> None:
-    """Write _URL style constant variables to a file bash scripts can load as env vars."""
+    """Write _URL style constant variables to URLS_ENV file so bash scripts can load as env vars."""
     url_vars = {k: v for k, v in vars(output_files).items() if k.endswith('URL') and not k.startswith('GH')}
 
     if not args.suppress_output:
@@ -190,8 +175,8 @@ def _verify_all_emails_were_printed(epstein_files: EpsteinFiles, already_printed
     logger.warning(f"Printed {len(already_printed_emails):,} emails of {len(email_ids_that_were_printed):,} unique file IDs.")
     missed_an_email = False
 
-    for email in epstein_files.emails:
-        if email.file_id not in email_ids_that_were_printed and not email.is_duplicate():
+    for email in epstein_files.non_duplicate_emails():
+        if email.file_id not in email_ids_that_were_printed:
             logger.warning(f"Failed to print {email.summary()}")
             missed_an_email = True
 

epstein_files/util/rich.py

@@ -33,6 +33,7 @@ GREY_NUMBERS = [58, 39, 39, 35, 30, 27, 23, 23, 19, 19, 15, 15, 15]
 DEFAULT_NAME_STYLE = 'gray46'
 INFO_STYLE = 'white dim italic'
 KEY_STYLE='honeydew2 bold'
+LAST_TIMESTAMP_STYLE='wheat4'
 SECTION_HEADER_STYLE = 'bold white on blue3'
 SOCIAL_MEDIA_LINK_STYLE = 'pale_turquoise4'
 SUBSTACK_POST_LINK_STYLE = 'bright_cyan'
@@ -79,10 +80,18 @@ console = Console(**CONSOLE_ARGS)
 highlighter = CONSOLE_ARGS['highlighter']
 
 
-def add_cols_to_table(table: Table, col_names: list[str]) -> None:
+def add_cols_to_table(table: Table, col_names: list[str | dict]) -> None:
     """Left most col will be left justified, rest are center justified."""
     for i, col in enumerate(col_names):
-        table.add_column(col, justify='left' if i == 0 else 'center')
+        if isinstance(col, dict):
+            col_name = col['name']
+            kwargs = col
+            del kwargs['name']
+        else:
+            col_name = col
+            kwargs = {}
+
+        table.add_column(col_name, justify='left' if i == 0 else 'center', **kwargs)
 
 
 def build_highlighter(pattern: str) -> EpsteinHighlighter:
@@ -93,8 +102,13 @@ def build_highlighter(pattern: str) -> EpsteinHighlighter:
     return TempHighlighter()
 
 
-def build_table(title: str | None, **kwargs) -> Table:
-    return Table(title=title, **{**DEFAULT_TABLE_KWARGS, **kwargs})
+def build_table(title: str | None, cols: list[str | dict] | None = None, **kwargs) -> Table:
+    table = Table(title=title, **{**DEFAULT_TABLE_KWARGS, **kwargs})
+
+    if cols:
+        add_cols_to_table(table, cols)
+
+    return table
 
 
 def join_texts(txts: list[Text], join: str = ' ', encloser: str = '', encloser_style: str = 'wheat4') -> Text:
@@ -129,6 +143,13 @@ def parenthesize(msg: str | Text, style: str = '') -> Text:
     return Text('(', style=style).append(txt).append(')')
 
 
+def print_all_files_page_link(epstein_files: 'EpsteinFiles') -> None:
+    markup_msg = link_markup(SITE_URLS[EMAIL], 'the other page', style='light_slate_grey bold')
+    txt = Text.from_markup(markup_msg).append(f' is uncurated and has all {len(epstein_files.other_files)}')
+    txt.append(f" unclassifiable files and {len(epstein_files.emails):,} emails")
+    print_centered(parenthesize(txt), style='dim')
+
+
 def print_author_header(msg: str, color: str | None, footer: str | None = None) -> None:
     txt = Text(msg, justify='center')
     color = color or 'white'
@@ -174,7 +195,20 @@ def print_header(epstein_files: 'EpsteinFiles') -> None:
 
     console.print(f"{not_optimized_msg}.\n", style='dim')
     print_page_title(width=TITLE_WIDTH)
-    print_other_site_link()
+    site_type: SiteType = EMAIL if args.all_emails else TEXT_MESSAGE
+    print_starred_header(f"This is the Epstein {site_type.title()}s site", num_spaces=4, num_stars=14)
+    other_site_type: SiteType = TEXT_MESSAGE if site_type == EMAIL else EMAIL
+    other_site_msg = "another site with" + (' all of' if other_site_type == EMAIL else '')
+    other_site_msg += f" Epstein's {other_site_type}s also generated by this code"
+
+    markup_msg = link_markup(SITE_URLS[other_site_type], other_site_msg, OTHER_SITE_LINK_STYLE)
+    print_centered(parenthesize(Text.from_markup(markup_msg)), style='bold')
+    word_count_link = link_text_obj(WORD_COUNT_URL, 'most frequently used words in the emails and texts', AUX_SITE_LINK_STYLE)
+    print_centered(parenthesize(word_count_link))
+    metadata_link = link_text_obj(JSON_METADATA_URL, 'author attribution explanations', AUX_SITE_LINK_STYLE)
+    print_centered(parenthesize(metadata_link))
+    json_link = link_text_obj(WORD_COUNT_URL, "epstein's json files", AUX_SITE_LINK_STYLE)
+    print_centered(parenthesize(json_link))
     _print_external_links()
     console.line()
     _print_abbreviations_table()
@@ -238,31 +272,6 @@ def print_numbered_list_of_emailers(_list: list[str | None], epstein_files = Non
     console.line()
 
 
-def print_other_site_link(is_header: bool = True) -> None:
-    """Print a link to the emails site if we're building text messages site and vice versa."""
-    site_type: SiteType = EMAIL if args.all_emails else TEXT_MESSAGE
-    link_style = OTHER_SITE_LINK_STYLE if is_header else 'light_slate_grey bold'
-
-    if is_header:
-        print_starred_header(f"This is the Epstein {site_type.title()}s site", num_spaces=4, num_stars=14)
-
-    other_site_type: SiteType = TEXT_MESSAGE if site_type == EMAIL else EMAIL
-    other_site_msg = "another site with" + (' all of' if other_site_type == EMAIL else '')
-    other_site_msg += f" Epstein's {other_site_type}s also generated by this code"
-    markup_msg = link_markup(SITE_URLS[other_site_type], other_site_msg, link_style)
-    print_centered(parenthesize(Text.from_markup(markup_msg)), style='bold')
-
-    if not is_header:
-        return
-
-    word_count_link = link_text_obj(WORD_COUNT_URL, 'most frequently used words in the emails and texts', AUX_SITE_LINK_STYLE)
-    print_centered(parenthesize(word_count_link))
-    metadata_link = link_text_obj(JSON_METADATA_URL, 'author attribution explanations', AUX_SITE_LINK_STYLE)
-    print_centered(parenthesize(metadata_link))
-    json_link = link_text_obj(WORD_COUNT_URL, "epstein's json files", AUX_SITE_LINK_STYLE)
-    print_centered(parenthesize(json_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)
     console.print(Align.center(vertically_pad(title_panel)))
@@ -285,7 +294,7 @@ def print_panel(msg: str, style: str = 'black on white', padding: tuple | None =
 def print_section_header(msg: str, style: str = SECTION_HEADER_STYLE, is_centered: bool = False) -> None:
     panel = Panel(Text(msg, justify='center'), expand=True, padding=(1, 1), style=style)
     panel = Align.center(panel) if is_centered else panel
-    console.print(Padding(panel, (3, 0, 1, 0)))
+    console.print(Padding(panel, (3, 2, 1, 2)))
 
 
 def print_starred_header(msg: str, num_stars: int = 7, num_spaces: int = 2, style: str = TITLE_STYLE) -> None:

epstein_files/util/word_count.py

@@ -198,7 +198,7 @@ def write_word_counts_html() -> None:
     word_count = WordCount()
 
     # Remove dupes, junk mail, and fwded articles from emails
-    emails = [e for e in epstein_files.emails if not (e.is_duplicate() or e.is_junk_mail() or e.is_fwded_article())]
+    emails = [e for e in epstein_files.non_duplicate_emails() if not (e.is_junk_mail() or e.is_fwded_article())]
 
     for email in emails:
         if specified_names and email.author not in specified_names:

{epstein_files-1.0.12.dist-info → epstein_files-1.0.14.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: epstein-files
-Version: 1.0.12
+Version: 1.0.14
 Summary: Tools for working with the Jeffrey Epstein documents released in November 2025.
 Home-page: https://michelcrypt4d4mus.github.io/epstein_text_messages/
 License: GPL-3.0-or-later
@@ -43,11 +43,12 @@ Description-Content-Type: text/markdown
 
 
 ## Usage
-
 #### Installation
 1. Requires you have a local copy of the OCR text files from the House Oversight document release in a directory `/path/to/epstein/ocr_txt_files`. You can download those OCR text files from [the Congressional Google Drive folder](https://drive.google.com/drive/folders/1ldncvdqIf6miiskDp_EDuGSDAaI_fJx8) (make sure you grab both the `001/` and `002/` folders).
 1. Use `poetry install` for easiest time installing. `pip install epstein-files` should also work, though `pipx install epstein-files` is usually better.
 
+
+#### Command Line Tools
 You need to set the `EPSTEIN_DOCS_DIR` environment variable with the path to the folder of files you just downloaded when running. You can either create a `.env` file modeled on [`.env.example`](./.env.example) (which will set it permanently) or you can run with:
 
 ```bash
@@ -63,7 +64,7 @@ epstein_generate
 # Search for a string:
 epstein_search Bannon
 # Or a regex:
-epstein_search '\bSteve\s*Bannon\b'
+epstein_search '\bSteve\s*Bannon|Jeffrey\s*Epstein\b'
 
 # Show a file with color highlighting of keywords:
 epstein_show 030999
@@ -82,6 +83,12 @@ epstein_diff 030999 020442
 The first time you run anything it will take a few minutes to fix all the janky OCR text, attribute the redacted emails, etc. After that things will be quick.
 Run `epstein_generate --help` for command line option assistance.
 
+**Optional:** There are a handful of emails that I extracted from the legal filings they were contained in. If you want to include these files in your local analysis you'll need to copy those files from the repo into your local document directory. Something like:
+
+```bash
+cp ./emails_extracted_from_legal_filings/*.txt "$EPSTEIN_DOCS_DIR"
+```
+
 
 #### As A Library
 ```python

epstein_files-1.0.14.dist-info/RECORD

@@ -0,0 +1,33 @@
+epstein_files/__init__.py,sha256=9mClHe3YkTWHulLOZIO2uaxYcff-krpEsc0vphYlwps,4975
+epstein_files/documents/communication.py,sha256=oqNsSDWe-N0jSmchIHxpihihzIWha-foFqMwKZlxyng,2057
+epstein_files/documents/document.py,sha256=1qL1V5VrrtI7uuuLvOw-JU83khRzzEURduuoh_pfghI,17114
+epstein_files/documents/email.py,sha256=eKfDGwbP1nYVfhozM0bSVf-5T-_l_k-0D_HVrQldm38,40904
+epstein_files/documents/emails/email_header.py,sha256=wkPfSLbmzkAeQwvhf0bAeFDLPbQT-EeG0v8vNNLYktM,7502
+epstein_files/documents/imessage/text_message.py,sha256=JPHlO-5Xgt5V5E9fDsgCEr3XOp7FyzlhIHmZhFNZcmg,2806
+epstein_files/documents/json_file.py,sha256=WcZW5NNqA67rHTdopbOGtup00muNaLlvrNgKb-K4zO8,1504
+epstein_files/documents/messenger_log.py,sha256=le-cj_ZBDUb_jnxP_F05KlnPjh0nDbgpPSWl4-HL-J4,6502
+epstein_files/documents/other_file.py,sha256=tJmEH3cGeEcRIZW5F-RuSJprYM_2r1jEbsF_kjSsA3E,9788
+epstein_files/epstein_files.py,sha256=aO2vntbA7DonSk96xB5us5VaDSpp79Mpu8eYJ_g_8Qo,20025
+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=smC9xsR9E9WECXi_ygK_6lrcEPZWZINLZrp_-BthUtA,10377
+epstein_files/util/constant/output_files.py,sha256=BkV4_gmdj46RfGy5SFYp6dgTty3FtlBth5YGmaGutls,1700
+epstein_files/util/constant/strings.py,sha256=etCmLZPuriVYDap4iG_0Nfd92P1cvRpK2R7ZpVAGDrw,1942
+epstein_files/util/constant/urls.py,sha256=EKzpL1np3cdyYvUlgv1zl9IJU8Vrg8jXpAu6YpM0_Bk,5283
+epstein_files/util/constants.py,sha256=W-4ibPdv8I62uNyoBUR7sWWiqUWWr2n414WnZNA_Who,112605
+epstein_files/util/data.py,sha256=JccGFZGiCGm7XtwpQTocIjGYOr6hTUpEPwHhjyW9Xnc,3164
+epstein_files/util/doc_cfg.py,sha256=uVUgZF7p5QMhbPEVauQWlRWQRhP0Imr5XaNQtMk0E7c,9053
+epstein_files/util/env.py,sha256=zg-TJ8ea6nTa0PyYgY97Zksre8au1ELv2i8vOizZt7Y,5645
+epstein_files/util/file_helper.py,sha256=tacTe1GcAnckPFvjMgxRRSLnFgr2aVIYsgfDR_C9uXk,2780
+epstein_files/util/highlighted_group.py,sha256=mNcdML0h88e2voEiUJxpgAVXHh3jQ8f0XV8qVdZMkBM,36550
+epstein_files/util/logging.py,sha256=fuREq06xUUI3DfCV2JE-8QM-sQKxpLDj0_AYFO6qR1M,1983
+epstein_files/util/output.py,sha256=bb_-jRoFNNMHXPSaptEXdkbqNDJYOpNSs6dkJKvi_HQ,7581
+epstein_files/util/rich.py,sha256=bQPlOEZ6qTpVF4yPnwmlBk4K11zH5RJjwbWeBIOUUGE,15191
+epstein_files/util/search_result.py,sha256=1fxe0KPBQXBk4dLfu6m0QXIzYfZCzvaSkWqvghJGzxY,567
+epstein_files/util/timer.py,sha256=8hxW4Y1JcTUfnBrHh7sL2pM9xu1sL4HFQM4CmmzTarU,837
+epstein_files/util/word_count.py,sha256=dTBNqX98sj7W46O0uit76I4hYBhm0bsY1Q7RHwWehiU,9349
+epstein_files-1.0.14.dist-info/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
+epstein_files-1.0.14.dist-info/METADATA,sha256=AUw8p6SQuvkJONsHErig0L--FyAU2y892DYpdpimqtg,5867
+epstein_files-1.0.14.dist-info/WHEEL,sha256=d2fvjOD7sXsVzChCqf0Ty0JbHKBaLYwDbGQDwQTnJ50,88
+epstein_files-1.0.14.dist-info/entry_points.txt,sha256=5qYgwAXpxegeAicD_rzda_trDRnUC51F5UVDpcZ7j6Q,240
+epstein_files-1.0.14.dist-info/RECORD,,

epstein_files-1.0.12.dist-info/RECORD

@@ -1,33 +0,0 @@
-epstein_files/__init__.py,sha256=qVFB7sS6XSlZX-ByyDwdbGSn2h06aoX2Mx8WcgRb-To,4710
-epstein_files/documents/communication.py,sha256=XapJlNfcaww3TpSkZIBE5c1Skqv_pFEFlIVi06V7k3E,2046
-epstein_files/documents/document.py,sha256=2FxyqWKROi7w2SmaQ493oGKekNvYAHSuv2YsDhPNQBU,16987
-epstein_files/documents/email.py,sha256=y8QTq349LWlm2LLUJ8rGcdkDbaGYJCV99wJytPcEMew,40587
-epstein_files/documents/emails/email_header.py,sha256=wkPfSLbmzkAeQwvhf0bAeFDLPbQT-EeG0v8vNNLYktM,7502
-epstein_files/documents/imessage/text_message.py,sha256=4gFvTfulj_Su10lNQl6Hq_p9ArTrSmn5pfC22YRJXjI,2794
-epstein_files/documents/json_file.py,sha256=tIYTwA3FYkwVZSpXvFYyUoH9m2sGYCD1U0ttamH6r1o,1306
-epstein_files/documents/messenger_log.py,sha256=yT4WQyTE_W6yelug_YGpBMRJ0YxWNtX4rKoEj8n5TMA,6260
-epstein_files/documents/other_file.py,sha256=pnl_q1o7ur3eeqGPwsYL2qbM3Y8O9LX6j6LbWnoxAiE,9939
-epstein_files/epstein_files.py,sha256=SaD4DJJ5tRxY97Ei4BdOgLzHQ9wrBVGrP64CSqdmk-w,18691
-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=CLWXrln8J-Dth6C-YF7Wdy7UoA8dybKJyqOLETrBeek,10284
-epstein_files/util/constant/output_files.py,sha256=BkV4_gmdj46RfGy5SFYp6dgTty3FtlBth5YGmaGutls,1700
-epstein_files/util/constant/strings.py,sha256=02DwbhAe8qBRq5HOUFx5FafXJ1P2-RJf9TCVu2b7UDQ,1932
-epstein_files/util/constant/urls.py,sha256=0IdCVVvXib0i-4TZFkVHoS4zCbjOBZWcr6NkGxsmQWM,4981
-epstein_files/util/constants.py,sha256=BpPRivoDYFI0uLU35kKpOdrSI6Rr9cmcrRj9-kANVrs,111834
-epstein_files/util/data.py,sha256=X3AutdW-ascIlE2bz1BtN0Bywqpe4OwYzJ-diEpfogI,2992
-epstein_files/util/doc_cfg.py,sha256=_f03qtA7qVbViHwqMXC4O5nfNbh90zDSq6El9Ior6f0,8996
-epstein_files/util/env.py,sha256=HnYcfHSNkwVJ_T75Woy43_OpDyxD0KHPj3GxcVx86N4,5751
-epstein_files/util/file_helper.py,sha256=tacTe1GcAnckPFvjMgxRRSLnFgr2aVIYsgfDR_C9uXk,2780
-epstein_files/util/highlighted_group.py,sha256=xrDLB05YUYGsU6vDvhvENMvIyjEz-9eb9xN-RjfCQbQ,36531
-epstein_files/util/logging.py,sha256=fuREq06xUUI3DfCV2JE-8QM-sQKxpLDj0_AYFO6qR1M,1983
-epstein_files/util/output.py,sha256=XcflgSOlzUGj6FsFaK6j4Dljld8A0h_uVV7ERcI_EYw,8120
-epstein_files/util/rich.py,sha256=8-4IA5bwPBdDPqkPdymq3zVKB9hfy3nrT7fUrN_XevY,14744
-epstein_files/util/search_result.py,sha256=1fxe0KPBQXBk4dLfu6m0QXIzYfZCzvaSkWqvghJGzxY,567
-epstein_files/util/timer.py,sha256=8hxW4Y1JcTUfnBrHh7sL2pM9xu1sL4HFQM4CmmzTarU,837
-epstein_files/util/word_count.py,sha256=8qBTuq3d0Q-3fwiuECKWi2RfL-KUiZD8TciwvfL0D_o,9353
-epstein_files-1.0.12.dist-info/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
-epstein_files-1.0.12.dist-info/METADATA,sha256=imTDdrHjWC-bWuw58SAyjYyiziZsqHkO7ODQUntw6YQ,5480
-epstein_files-1.0.12.dist-info/WHEEL,sha256=d2fvjOD7sXsVzChCqf0Ty0JbHKBaLYwDbGQDwQTnJ50,88
-epstein_files-1.0.12.dist-info/entry_points.txt,sha256=5qYgwAXpxegeAicD_rzda_trDRnUC51F5UVDpcZ7j6Q,240
-epstein_files-1.0.12.dist-info/RECORD,,