# Export dump files as JSON

**URL:** <https://discourse.mitmproxy.org/t/export-dump-files-as-json/166>\
**Category:** Uncategorized\
**Created:** [September 9, 2016, 4:13pm UTC](https://discourse.mitmproxy.org/t/export-dump-files-as-json/166 "2016-09-09T16:13:30Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![xrd](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.mitmproxy.org/xrd/32/57_2.png) [@xrd](https://discourse.mitmproxy.org/u/xrd)\
**Post date:** [September 9, 2016, 4:13pm UTC](https://discourse.mitmproxy.org/t/export-dump-files-as-json/166/1 "2016-09-09T16:13:30Z")

</div>

I want to export my mitmproxy dump files (captured using -w switch) to JSON.

I wrote a script which does this, more or less, here:

```python
from mitmproxy import flow
import sys
import jsonpickle

items = []
with open(sys.argv[1], "rb") as logfile:
    freader = flow.FlowReader(logfile)
    try:
        for f in freader.stream():
	    items.append( f )
    except Exception as e:
        print("Flow file corrupted: {}".format(e))

    print jsonpickle.encode( items )

```

This works, but (as expected) is dumping out the python objects. The output looks like this:

```json
[[{"py/object":"mitmproxy.models.http.HTTPFlow","server_conn":{"py/object":"mitmproxy.models.connections.ServerConnection","server_certs":[],"via":null,"protocol":null,"timestamp_tcp_setup":1473286806.433876,"cert":{"py/object":"netlib.certutils.SSLCert","x509":{"py/object":"OpenSSL.crypto.X509","_x509":{"py/object":"_cffi_backend.CDataGCP"}}},"timestamp_ssl_setup":1473286806.68306,"_TCPClient__source_address":{"py/object":"netlib.tcp.Address","family":2,"address":{"py/tuple":["1...

```

I’d love to find a better way to dump the files which gets me closer to the raw HTTP request/response itself and less of the internal python details. Perhaps go to HAR file, and then convert to JSON? I’m not much of a python programmer, so I’m unsure how to get started inspecting the objects and pulling out the important information.

My goal is to have:

- the request: headers and body
- the response: headers and body
- host and path.
- maybe timing information?

Any suggestions?

---

<div class="post-metadata">

**Author:** ![mhils](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.mitmproxy.org/mhils/32/7_2.png) [@mhils](https://discourse.mitmproxy.org/u/mhils)\
**Post date:** [September 10, 2016, 12:02am UTC](https://discourse.mitmproxy.org/t/export-dump-files-as-json/166/2 "2016-09-10T00:02:39Z")

</div>

Hi @xrd,

I believe your semantic mistake is using jsonpickle, which would serialize whole objects - you probably just want to use the `json` module from the standard library. Then you can either use `json.dumps(flow.get_state())` or construct the dictionary yourself and then `json.dumps()` that one. [https://github.com/mitmproxy/mitmproxy/blob/master/examples/har\_dump.py](https://github.com/mitmproxy/mitmproxy/blob/master/examples/har_dump.py) shows that on a more advanced level!