Integrate session data for file context in sidekick function

Resolves #65

The sidekick function now utilizes session data to determine file context. If files from the last session are available, they are used for context. Otherwise, recent git commits and current changes are used. This change also includes updating the session data with the new list of files whenever the files change. This enhancement should provide a more consistent and relevant context for the sidekick function.

aicodebot/cli.py

@@ -1,7 +1,7 @@
 from aicodebot import version as aicodebot_version
 from aicodebot.agents import SidekickAgent
 from aicodebot.coder import CREATIVE_TEMPERATURE, DEFAULT_MAX_TOKENS, Coder
-from aicodebot.config import get_config_file, get_local_data_dir, read_config
+from aicodebot.config import Session, get_config_file, get_local_data_dir, read_config
 from aicodebot.helpers import create_and_write_file, exec_and_get_output, logger
 from aicodebot.input import Chat, SidekickCompleter
 from aicodebot.learn import load_documents_from_repo, store_documents
@@ -474,8 +474,14 @@ def sidekick(request, verbose, no_files, max_file_tokens, files):  # noqa: PLR09
             )
     elif not no_files:
         # Determine which files to use for context automagically, with git
-        console.print("Using recent git commits and current changes for context.")
-        files = Coder.auto_file_context(file_context_limit, max_file_tokens)
+        session_data = Session.read()
+        if session_data.get("files"):
+            console.print("Using files from the last session for context.")
+            files = session_data["files"]
+        else:
+            console.print("Using recent git commits and current changes for context.")
+            files = Coder.auto_file_context(file_context_limit, max_file_tokens)
+
         context = generate_files_context(files)
         file_token_size = Coder.get_token_length(context)
     else:
@@ -539,7 +545,11 @@ def calc_response_token_size(files):
         # Update the context for the new list of files
         context = generate_files_context(chat.files)
         languages = ",".join(Coder.identify_languages(chat.files))
-        completer.files = chat.files
+        if completer.files != chat.files:
+            completer.files = chat.files
+            session_data = Session.read()
+            session_data["files"] = list(chat.files)
+            Session.write(session_data)
 
         if parsed_human_input == chat.CONTINUE:
             continue