Autopsy  4.21.0
Graphical digital forensics platform for The Sleuth Kit and other tools.
Python Tutorial #2: Writing a Data Source Ingest Module

In the first tutorial we built a basic Python Autopsy module that looked for big and round files. In this tutorial we're going to make two data source ingest modules. The first focuses on finding SQLite databases and parsing them, and the second focuses on running a command line tool on a disk image.

The main difference from the first tutorial, which focused on file ingest modules, is that these are data source ingest modules. Data source-ingest modules are given a reference to a data source and the module needs to find the files to analyze, whereas file-level ingest modules are given a reference to each file in the data source.


This post assumes you've read the first tutorial. That means that you should know why it is better to write an Autopsy module versus a stand-alone tool, and what you need to set up (Autopsy installed, text editor, etc.). You may also recall the limitations (and benefits) of data source ingest modules. The most notable difference between them is that data source-ingest modules may not have access to carved files or files that are inside of ZIP files. For our example in this post, we are looking for a SQLite database with a specific name, and it will not be inside of a ZIP file, so data source ingest modules are the most efficient and will get us results faster.

The other assumption is that you know something about SQL queries. We have some example queries below and we don't go into detail about how they work.

Getting Started

Making Your Module Folder

We'll start by making our module folder. As we learned in the first tutorial, every Python module in Autopsy gets its own folder. To find out where you should put your Python module, launch Autopsy and choose the Tools->Python Plugins menu item. That will open a subfolder in your AppData folder, such as "C:\Users\JDoe\AppData\Roaming\Autopsy\python_modules".

Make a folder inside of there to store your module. Call it "DemoScript2". Copy the sample file from github into the this new folder and rename it to

Writing The Script

We are going to write a script that:

Open the script in your favorite text editor. The sample Autopsy Python modules all have TODO entries in them to let you know what you should change. The below steps jump from one TODO to the next.

  1. Factory Class Name: The first thing to do is rename the sample class name from "SampleJythonDataSourceIngestModuleFactory" to "ContactsDbIngestModuleFactory". In the sample module, there are several uses of this class name, so you should search and replace for these strings.
  2. Name and Description: The next TODO entries are for names and descriptions. These are shown to users. For this example, we'll name it "Contacts Db Analyzer". The description can be anything you want. Note that Autopsy requires that modules have unique names, so don't make it too generic.
  3. Ingest Module Class Name: The next thing to do is rename the ingest module class from "SampleJythonDataSourceIngestModule" to "ContactsDbIngestModule". Our usual naming convention is that this class is the same as the factory class with "Factory" removed from the end. There are a couple of places where this name is used, so do a search and replace in your code.
  4. startUp() method: The startUp() method is where each module initializes. For our example, we don't need to do anything special in here except save a reference to the passed in context object. This is used later on to see if the module has been cancelled.
  5. process() method: This is where we do our analysis and we'll focus on this more in the next section.

That's it. In the file-level ingest module, we had a shutdown() method, but we do not need that with data source-level ingest modules. When their process method is finished, it can shut itself down. The process() method will be called only once.

The process() Method

The process method in a data source-level ingest module is passed in reference to the data source as a Content object and a Progress Bar class to update our progress.

For this tutorial, you can start by deleting the contents of the existing process() method in the sample module. The full source code is linked to at the end of this blog and shows more detail about a fully fledged module. We'll just cover the analytics in the blog.

Getting Files

Because data source-level ingest modules are not passed in specific files to analyze, nearly all of these types of modules will need to use the service to find relevant files. Check out the methods on that class to see the different ways that you can find files.

NOTE: See the Running Executables section for an example of when you simply want to run a command line tool on a disk image instead of querying for files to analyze.

For our example, we want to find all files named "contacts.db". The class contains several findFiles() methods to help. You can search for all files with a given name or files with a given name in a particular folder. You can also use SQL syntax to match file patterns, such as "%.jpg" to find all files with a JPEG extension.

Our example needs these two lines to get the FileManager for the current case and to find the files.

fileManager = Case.getCurrentCase().getServices().getFileManager()
files = fileManager.findFiles(dataSource, "contacts.db")

findFiles() returns a list of AbstractFile objects. This gives you access to the file's metadata and content.

For our example, we are going to open these SQLite files. That means that we need to save them to disk. This is less than ideal because it wastes time writing the data to disk and then reading it back in, but it is the only option with many libraries. If you are doing some other type analysis on the content, then you do not need to write it to disk. You can read directly from the AbstractFile (see the sample modules for specific code to do this).

The org.sleuthkit.autopsy.datamodel.ContentUtils class provides a utility to save file content to disk. We'll make a path in the temp folder of our case directory. To prevent naming collisions, we'll name the file based on its unique ID. The following two lines save the file to lclDbPath.

lclDbPath = os.path.join(Case.getCurrentCase().getTempDirectory(), str(file.getId()) + ".db")
ContentUtils.writeToFile(file, File(lclDbPath))

Analyzing SQLite

Next, we need to open the SQLite database. We are going to use the Java JDBC infrastructure for this. JDBC is Java's generic way of dealing with different types of databases. To open the database, we do this:

dbConn = DriverManager.getConnection("jdbc:sqlite:%s"  % lclDbPath)

With our connection in hand, we can do some queries. In our sample database, we have a single table named "contacts", which has columns for name, email, and phone. We first start by querying for all rows in our simple table:

stmt = dbConn.createStatement()
resultSet = stmt.executeQuery("SELECT * FROM contacts")

For each row, we are going to get the values for the name, e-mail, and phone number and make a TSK_CONTACT artifact. Recall from the first tutorial that posting artifacts to the blackboard allows modules to communicate with each other and also allows you to easily display data to the user. The TSK_CONTACT artifact is for storing contact information. The artifact catalog shows that TSK_CONTACT is a data artifact, so we will be using the newDataArtifact() method to create each one.

The basic approach in our example is to make an artifact of a given type (TSK_CONTACT) and have it be associated with the database it came from. We then make attributes for the name, email, and phone. The following code does this for each row in the database:

                name  = resultSet.getString("name")
                email = resultSet.getString("email")
                phone = resultSet.getString("phone")
        except SQLException as e:
                self.log(Level.INFO, "Error getting values from contacts table (" + e.getMessage() + ")")
        # Make an artifact on the blackboard, TSK_CONTACT and give it attributes for each of the fields
        art = file.newDataArtifact(BlackboardArtifact.Type.TSK_CONTACT, Arrays.asList(
                                                        ContactsDbIngestModuleFactory.moduleName, name),
                                                        ContactsDbIngestModuleFactory.moduleName, email),
                                                        ContactsDbIngestModuleFactory.moduleName, phone)

That's it. We've just found the databases, queried them, and made artifacts for the user to see. There are some final things though. First, we should fire off an event so that the UI updates and refreshes with the new artifacts. We can fire just one event after each database is parsed (or you could fire one for each artifact - it's up to you).

  BlackboardArtifact.ARTIFACT_TYPE.TSK_CONTACT, None))

And the final thing is to clean up. We should close the database connections and delete our temporary file.


The final version of can be found on github.


Data source-level ingest modules can run for quite some time. Therefore, data source-level ingest modules should do some additional things that file-level ingest modules do not need to.

Debugging and Development Tips

You can find the full file along with a small sample database on github. To use the database, add it as a logical file and run your module on it.

Whenever you have syntax errors or other errors in your script, you will get some form of dialog from Autopsy when you try to run ingest modules. If that happens, fix the problem and run ingest modules again. You don't need to restart Autopsy each time!

The sample module has some log statements in there to help debug what is going on since we don't know of better ways to debug the scripts while running in Autopsy.

Running Executables

While the above example outlined using the FileManager to find files to analyze, the other common use of data source-level ingest modules is to wrap a command line tool that takes a disk image as input. A sample program ( that does that can be found on github. I'll cover the big topics of that program in this section. There are more details in the script about error checking and such.

Finding The Executable

To write this kind of data source-level ingest module, put the executable in your module's folder (the DemoScript2 folder we previously made). Use "__file__" to get the path to where your script is and then use some os.path methods to get to the executable in the same folder.

path_to_exe = os.path.join(os.path.dirname(os.path.abspath(__file__)), "img_stat.exe")

In our sample program, we do this and verify we can find it in the startup() method so that if we don't, then ingest never starts.

Running The Executable

Data sources can be disk images, but they can also be a folder of files. We only want to run our executable on a disk image. So, verify that:

if not isinstance(dataSource, Image):
   self.log(Level.INFO, "Ignoring data source.  Not an image")
   return IngestModule.ProcessResult.OK 

You can get the path to the disk image using dataSource.getPaths().

Once you have the EXE and the disk image, you can use the various subprocess methods to run them.

Showing the User Results

After the command line tool runs, you have the option of either showing the user the raw output of the tool or parsing it into individual artifacts. Refer to previous sections of this tutorial and the previous tutorial for making artifacts. If you want to simply show the user the output of the tool, then save the output to the Reports folder in the Case directory:

reportPath = os.path.join(Case.getCurrentCase().getCaseDirectory(),
    "Reports", "img_stat-" + str(dataSource.getId()) + ".txt") 

Then you can add the report to the case so that it shows up in the tree in the main UI panel.

Case.getCurrentCase().addReport(reportPath, "Run EXE", "img_stat output")


Data source-level ingest modules allow you to query for a subset of files by name or to run on an entire disk image. This tutorial has shown an example of both use cases and shown how to use SQLite in Jython.

Copyright © 2012-2022 Basis Technology. Generated on: Tue Feb 6 2024
This work is licensed under a Creative Commons Attribution-Share Alike 3.0 United States License.