feat: add group filter option by sachitv · Pull Request #639 · ReagentX/imessage-exporter

sachitv · 2025-11-24T07:50:37Z

This is based on the discussion here: #638

Copilot

Pull request overview

This pull request adds a new --group-filter (-g) option to filter exported group conversations by their display names. The feature works independently or in conjunction with the existing --conversation-filter option, using set intersection when both are specified.

Key changes:

New CLI option --group-filter for filtering group chats by name
Runtime logic to resolve group names to chat IDs and intersect with existing filters
Comprehensive test coverage for various filter scenarios
Documentation updates explaining the intersection behavior

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
imessage-exporter/src/main.rs	Adds call to `resolve_filtered_groups()` after handle resolution
imessage-exporter/src/app/options.rs	Defines new CLI option, integrates into validation logic, updates test data
imessage-exporter/src/app/runtime.rs	Implements group filter resolution logic, validation, and comprehensive tests
imessage-exporter/src/exporters/txt.rs	Fixes hardcoded home directory path in test to use environment variable
imessage-exporter/src/exporters/html.rs	Fixes hardcoded home directory path in test to use environment variable
imessage-exporter/README.md	Documents new option and intersection behavior with conversation filter

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-01-09T17:32:42Z

+                None => {
+                    self.options
+                        .query_context
+                        .set_selected_chat_ids(matched_chat_ids.clone());


The .clone() call on matched_chat_ids is necessary here since the variable is moved into set_selected_chat_ids. However, you could avoid this clone by restructuring the logic. Instead of matching on a reference and then calling the method separately, you could consume matched_chat_ids directly in both branches. For example, use if let Some(existing) = self.options.query_context.selected_chat_ids.take() and then set the result in both branches.

Copilot is right here, in the None branch this clones a value that is not used afterward; pass it by value instead.

Copilot · 2026-01-09T17:32:43Z

+            "Sticker from Sample Contact: {sticker_path} (App: Free People) from Sample Contact",
+            sticker_path = sticker_path


The sticker_path named parameter in the format! macro is redundant. The variable is already captured by name in the format string. Remove the explicit sticker_path = sticker_path parameter.

Suggested change

"Sticker from Sample Contact: {sticker_path} (App: Free People) from Sample Contact",

sticker_path = sticker_path

"Sticker from Sample Contact: {sticker_path} (App: Free People) from Sample Contact"

Copilot · 2026-01-09T17:32:43Z

+            "<img src=\"{sticker_path}\" loading=\"lazy\">\n<div class=\"sticker_name\">App: Free People</div> <div class=\"sticker_tapback\">&nbsp;by Sample Contact</div>",
+            sticker_path = sticker_path


The sticker_path named parameter in the format! macro is redundant. The variable is already captured by name in the format string. Remove the explicit sticker_path = sticker_path parameter.

Suggested change

"<img src=\"{sticker_path}\" loading=\"lazy\">\n<div class=\"sticker_name\">App: Free People</div> <div class=\"sticker_tapback\"> by Sample Contact</div>",

sticker_path = sticker_path

"<img src=\"{sticker_path}\" loading=\"lazy\">\n<div class=\"sticker_name\">App: Free People</div> <div class=\"sticker_tapback\"> by Sample Contact</div>"

ReagentX · 2026-01-09T17:33:37Z

Ignore copilot for now; I haven not yet decided if I want to include this feature yet.

sachitv · 2026-01-23T16:26:31Z

@ReagentX any update on this yet?

ReagentX · 2026-01-23T16:31:56Z

Not yet

ReagentX · 2026-05-27T16:12:22Z

Thanks for the update. I haven't lost track of this one.

ReagentX

The overall approach is right, and the tests cover the cases that matter, we just need a few refinements to land this.

ReagentX · 2026-06-06T19:07:24Z

+                .chatrooms
+                .iter()
+                .filter_map(|(&chat_id, chat)| {
+                    let chat_name = chat.name();


Chat::name() is the wrong source here. It falls back to chat_identifier when there is no custom name:

imessage-exporter/imessage-database/src/tables/chat.rs

Lines 120 to 127 in 9423e61

/// Generate a name for a chat, falling back to the default if a custom one is not set

#[must_use]

pub fn name(&self) -> &str {

match self.display_name() {

Some(name) => name,

None => &self.chat_identifier,

}

}

This means --group-filter does not actually filter on group names. For a 1:1 conversation, chat_identifier is the phone number or email, so -g "5558675309" matches a DM. For a group with no custom name, the only thing to match against is the opaque chatNNN… identifier, which no user would type. In practice the flag currently matches "any chat whose identifier or name contains the substring," which is not the goal of these changes.

Match on display_name() instead: a present display name is, in practice, the definition of a chat the user named, which is what we want this flag to select.

imessage-exporter/imessage-database/src/tables/chat.rs

Lines 129 to 141 in 9423e61

/// Get the current display name for the chat, if it exists.

#[must_use]

pub fn display_name(&self) -> Option<&str> {

match &self.display_name {

Some(name) => {

if !name.is_empty() {

return Some(name.as_str());

}

None

}

None => None,

}

}

ReagentX · 2026-06-06T19:11:52Z

    /// App configuration options
    pub options: Options,
+    /// Tracks whether the group filter matched any chats
+    pub group_filter_matches: bool,


This field is transient validation state living on a long-lived struct. resolve_filtered_groups writes it, start reads it, and the two only line up because of an undocumented requirement that resolution runs first. That is a coupling waiting to break: reorder the calls and the guard silently stops working.

I would rather resolve_filtered_groups return its outcome as Result<(), RuntimeError> and let main.rs, where the filters are already resolved, surface the error. This field then field disappears, there is nothing to keep in sync, and the ordering requirement becomes a compile-time argument rather than a convention.

ReagentX · 2026-06-06T19:13:48Z

+        app::{contacts::Name, export_type::ExportType, runtime::filename_tests::fake_chat},
+    };
+
+    /// Verify that when the group filter matches no chats, the conversation filter is still honored.


This comment does not match runtime behavior. When group_filter is set and group_filter_matches is false, start() returns InvalidOptions and the export aborts. The preserved selected_chat_ids is never used.

The test verifies internal state that the runtime immediately discards, which is the opposite of what happens. Either the export should proceed on the conversation filter in this case, or the test's premise should be corrected.

ReagentX · 2026-06-06T19:15:28Z


+            if let Some(filters) = &self.options.group_filter && !self.group_filter_matches {
+                return Err(RuntimeError::InvalidOptions(format!(
+                    "Selected group filter `{filters}` does not match any chats!"


This log fires in two distinct situations: the group matched nothing, and the group matched chats that do not intersect --conversation-filter. Those are different problems: the second one is not "no match", rather, both filters matched, they just do not overlap.

ReagentX · 2026-06-06T19:16:30Z

+                .collect();
+
+            if matched_chat_ids.is_empty() {
+                self.group_filter_matches = false;


Redundant as self.group_filter_matches is set to False at the top of the function.

ReagentX · 2026-06-06T19:16:38Z

+                        .collect::<BTreeSet<i32>>();
+
+                    if intersection.is_empty() {
+                        self.group_filter_matches = false;


Redundant as self.group_filter_matches is set to False at the top of the function.

ReagentX · 2026-06-06T19:17:10Z

+                None => {
+                    self.options
+                        .query_context
+                        .set_selected_chat_ids(matched_chat_ids.clone());


Copilot is right here, in the None branch this clones a value that is not used afterward; pass it by value instead.

ReagentX · 2026-06-06T19:17:57Z

+
+            if matched_chat_ids.is_empty() {
+                self.group_filter_matches = false;
+                return;


This branch sets the flag and returns, which is the same shape as the no-overlap branch above it. If this function returns a Result (see the comment on group_filter_matches), both collapse into one error path.

ReagentX · 2026-06-06T19:18:39Z

+        To provide multiple group names, use a comma-separated string
+        Example: `-g "Family Chat,Work Group"`
+
+Note: When both `--conversation-filter` and `--group-filter` are specified, only conversations that match both filters (intersection) will be exported.


The note should be indented as it would be displayed in the shell when you actually run -h, or it should exist elsewhere. Notes do not belong in this section as it is the literal output of -h.

ReagentX mentioned this pull request Dec 2, 2025

Failing tests on Linux #640

Closed

ReagentX self-assigned this Dec 6, 2025

ReagentX added new feature Requires creating a new feature crate: cli Related to the CLI crate labels Dec 6, 2025

ReagentX added this to 3.3: Chaparral Lotus Dec 11, 2025

ReagentX moved this to Todo in 3.3: Chaparral Lotus Dec 11, 2025

Copilot AI review requested due to automatic review settings January 9, 2026 17:26

Copilot started reviewing on behalf of ReagentX January 9, 2026 17:26 View session

Copilot AI reviewed Jan 9, 2026

View reviewed changes

ReagentX removed this from 3.3: Chaparral Lotus Jan 15, 2026

divinemystery111-dev approved these changes May 26, 2026

View reviewed changes

feat: add group filter option

1c7bea0

sachitv force-pushed the group_filter branch from 1cb864d to 1c7bea0 Compare May 27, 2026 15:41

ReagentX added this to 4.2: Stream Orchid May 29, 2026

ReagentX moved this to Todo in 4.2: Stream Orchid May 29, 2026

ReagentX requested changes Jun 6, 2026

View reviewed changes

ReagentX moved this from Todo to In Progress in 4.2: Stream Orchid Jun 6, 2026

ReagentX removed this from 4.2: Stream Orchid Jun 18, 2026

ReagentX assigned sachitv and unassigned ReagentX Jul 24, 2026

		"Sticker from Sample Contact: {sticker_path} (App: Free People) from Sample Contact",
		sticker_path = sticker_path

		"<img src=\"{sticker_path}\" loading=\"lazy\">\n<div class=\"sticker_name\">App: Free People</div> <div class=\"sticker_tapback\"> by Sample Contact</div>",
		sticker_path = sticker_path

	/// Generate a name for a chat, falling back to the default if a custom one is not set
	#[must_use]
	pub fn name(&self) -> &str {
	match self.display_name() {
	Some(name) => name,
	None => &self.chat_identifier,
	}
	}

	/// Get the current display name for the chat, if it exists.
	#[must_use]
	pub fn display_name(&self) -> Option<&str> {
	match &self.display_name {
	Some(name) => {
	if !name.is_empty() {
	return Some(name.as_str());
	}
	None
	}
	None => None,
	}
	}

Conversation

sachitv commented Nov 24, 2025

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Copilot AI Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

ReagentX commented Jan 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

sachitv commented Jan 23, 2026

Uh oh!

ReagentX commented Jan 23, 2026

Uh oh!

ReagentX commented May 27, 2026

Uh oh!

ReagentX left a comment

Choose a reason for hiding this comment

Uh oh!

ReagentX Jun 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

ReagentX commented Jan 9, 2026 •

edited

Loading

ReagentX Jun 6, 2026 •

edited

Loading