1 files changed, 87 insertions, 60 deletions
diff --git a/src/server/ApiManagers/AssistantManager.ts b/src/server/ApiManagers/AssistantManager.ts
index 224d47d3b..42e544e1d 100644
--- a/src/server/ApiManagers/AssistantManager.ts
+++ b/src/server/ApiManagers/AssistantManager.ts
@@ -442,69 +442,96 @@ export default class AssistantManager extends ApiManager {
 }
 
 function spawnPythonProcess(jobId: string, file_name: string, file_data: string) {
-    const pythonPath =
-        process.platform === 'win32'
-            ? path.join(__dirname, '../chunker/venv/Scripts/python') // Windows path
-            : path.join(__dirname, '../chunker/venv/bin/python3'); // Linux/Mac path
-
-    const pythonProcess = spawn(pythonPath, [
-        path.join(__dirname, '../chunker/pdf_chunker.py'), // Correct path to Python script
-        jobId,
-        file_name,
-        file_data,
-    ]);
-
-    let pythonOutput = ''; // Accumulate stdout data
-    let stderrOutput = ''; // For stderr logs and progress
-
-    // Handle stdout data (final result in JSON format)
-    pythonProcess.stdout.on('data', data => {
-        pythonOutput += data.toString(); // Accumulate data from stdout
-    });
-
-    // Handle stderr (progress logs or errors)
-    pythonProcess.stderr.on('data', data => {
-        stderrOutput += data.toString();
-        const lines = stderrOutput.split('\n');
-        lines.forEach(line => {
-            if (line.trim()) {
-                try {
-                    // Progress and warnings are printed as JSON to stderr
-                    const parsedOutput = JSON.parse(line);
-
-                    // Handle progress updates
-                    if (parsedOutput.job_id && parsedOutput.progress !== undefined) {
-                        jobProgress[parsedOutput.job_id] = {
-                            step: parsedOutput.step,
-                            progress: parsedOutput.progress,
-                        };
-                    } else if (parsedOutput.progress !== undefined) {
-                        jobProgress[jobId] = {
-                            step: parsedOutput.step,
-                            progress: parsedOutput.progress,
-                        };
+    const venvPath = path.join(__dirname, '../chunker/venv');
+    const requirementsPath = path.join(__dirname, '../chunker/requirements.txt');
+    const pythonScriptPath = path.join(__dirname, '../chunker/pdf_chunker.py');
+
+    // Check if venv exists
+    if (!fs.existsSync(venvPath)) {
+        console.log('Virtual environment not found. Creating and setting up...');
+
+        // Create venv
+        const createVenvProcess = spawn('python', ['-m', 'venv', venvPath]);
+
+        createVenvProcess.on('close', code => {
+            if (code !== 0) {
+                console.error(`Failed to create virtual environment. Exit code: ${code}`);
+                return;
+            }
+
+            console.log('Virtual environment created. Installing requirements...');
+
+            // Determine the pip path based on the OS
+            const pipPath = process.platform === 'win32' ? path.join(venvPath, 'Scripts', 'pip') : path.join(venvPath, 'bin', 'pip');
+
+            // Install requirements
+            const installRequirementsProcess = spawn(pipPath, ['install', '-r', requirementsPath]);
+
+            installRequirementsProcess.on('close', code => {
+                if (code !== 0) {
+                    console.error(`Failed to install requirements. Exit code: ${code}`);
+                    return;
+                }
+
+                console.log('Requirements installed. Running Python script...');
+                runPythonScript();
+            });
+        });
+    } else {
+        console.log('Virtual environment found. Running Python script...');
+        runPythonScript();
+    }
+
+    function runPythonScript() {
+        const pythonPath = process.platform === 'win32' ? path.join(venvPath, 'Scripts', 'python') : path.join(venvPath, 'bin', 'python3');
+
+        const pythonProcess = spawn(pythonPath, [pythonScriptPath, jobId, file_name, file_data]);
+
+        let pythonOutput = '';
+        let stderrOutput = '';
+
+        pythonProcess.stdout.on('data', data => {
+            pythonOutput += data.toString();
+        });
+
+        pythonProcess.stderr.on('data', data => {
+            stderrOutput += data.toString();
+            const lines = stderrOutput.split('\n');
+            lines.forEach(line => {
+                if (line.trim()) {
+                    try {
+                        const parsedOutput = JSON.parse(line);
+                        if (parsedOutput.job_id && parsedOutput.progress !== undefined) {
+                            jobProgress[parsedOutput.job_id] = {
+                                step: parsedOutput.step,
+                                progress: parsedOutput.progress,
+                            };
+                        } else if (parsedOutput.progress !== undefined) {
+                            jobProgress[jobId] = {
+                                step: parsedOutput.step,
+                                progress: parsedOutput.progress,
+                            };
+                        }
+                    } catch (err) {
+                        console.error('Progress log from Python:', line);
                     }
-                } catch (err) {
-                    console.error('Progress log from Python:', line);
                 }
-            }
+            });
         });
-    });
 
-    // Handle process exit
-    pythonProcess.on('close', code => {
-        if (code === 0) {
-            // Parse final JSON output (stdout)
-            try {
-                const finalResult = JSON.parse(pythonOutput); // Parse JSON from stdout
-                jobResults[jobId] = finalResult;
-                jobProgress[jobId] = { step: 'Complete', progress: 100 };
-            } catch (err) {
-                console.error('Error parsing final JSON result:', err);
+        pythonProcess.on('close', code => {
+            if (code === 0) {
+                try {
+                    const finalResult = JSON.parse(pythonOutput);
+                    jobResults[jobId] = finalResult;
+                    jobProgress[jobId] = { step: 'Complete', progress: 100 };
+                } catch (err) {
+                    console.error('Error parsing final JSON result:', err);
+                }
+            } else {
+                console.error(`Python process exited with code ${code}`);
+                jobResults[jobId] = { error: 'Python process failed' };
             }
-        } else {
-            console.error(`Python process exited with code ${code}`);
-            jobResults[jobId] = { error: 'Python process failed' };
-        }
-    });
+        });
+    }
 }