How to collect massive replay dem file efficiently?

[RuiLiFeng]

Hi all,

I want to collect many dota2 replay data for research purposes. However, currently, I cannot find an efficient way to do this.

First I question api.opendota.com/{matchpool}Matches (matchpool=pro or public) to get the matchids，then I visit api.opendota.com/matches/matchid to find whether it has a replay_url.

However in this procedure, about merely 1/20 matchid have replay_urls, which make my collection of replay dem file extremely expensive, cost nearly 0.3$ for one replay.

How can accelerate this process and make it much cheaper?